Skip to content
ConoceNivel 0 · by MetodologIAGo to MetodologIA
Your position · Module 3 of 4Amplified workView module in Resources
Module 03 · Prompts

Library: direct co-creation

Choose a task, observe your practice, compare sources, co-create the Mini-SOP, and test one case under human review.

How to use this library14 · Prompt structure levels
10 working prompts

Move from the situation to the outcome through a verifiable sequence.

Prompt library with levels and expected evidence.
WorkflowFrom a personal task to a tested Mini-SOP

Ten prompts connect map, standard, procedure, test, AAR, and transfer.

Independent use: Each works alone with its required input.
LearnChoose the workflowBound a useful workflow.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When starting from a task or workflow that is not yet bounded.
Example
Case: prepare and review a scope proposal. Start: brief received. End: version ready for a decision. Risk: citing an unverified source. Owner and reviewer: you, at separate moments.
Evidence
Has scope and owner.; A case exists.
Limit
Does not execute.
Receivesworkflow context
ProducesWorkflow brief
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <WORKFLOW>work with a start, finish and outputExample: prepare a proposal · Required
  • <BOUNDARIES>workflow start, finish and exclusionsExample: from brief to approval · Required
  • <RECORDS>observed calendar, task and delivery dataExample: five-day calendar · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Bound a useful workflow. The final decision and evidence remain under your responsibility.Data: WORKFLOW=[WORKFLOW · work with a start, finish and output · e.g.: prepare a proposal]; BOUNDARIES=[BOUNDARIES · workflow start, finish and exclusions · e.g.: from brief to approval]; RECORDS=[RECORDS · observed calendar, task and delivery data · e.g.: five-day calendar]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Light SIPOC; Definition of Done; Human-in-the-loop.Order: Define purpose and outcome. → Return the artifact, next step and any open gap.Deliver: Workflow brief; Has scope and owner.Boundary: Do not invent my context.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Bound a useful workflow. The final decision and evidence remain under your responsibility.Data: WORKFLOW=from receiving the brief to delivering a reviewed proposal; BOUNDARIES=starts with approved brief, ends with reviewed proposal and excludes negotiation; RECORDS=five-day calendar with 17 blocks, 12 tasks and one delivered proposal; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Case: prepare and review a scope proposal. Start: brief received. End: version ready for a decision. Risk: citing an unverified source. Owner and reviewer: you, at separate moments.Apply: Light SIPOC; Definition of Done; Human-in-the-loop.Order: Define purpose and outcome. → Return the artifact, next step and any open gap.Deliver: Workflow brief; Has scope and owner.Boundary: Do not invent my context.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSWORKFLOW = [WORKFLOW · work with a start, finish and output · e.g.: prepare a proposal]BOUNDARIES = [BOUNDARIES · workflow start, finish and exclusions · e.g.: from brief to approval]RECORDS = [RECORDS · observed calendar, task and delivery data · e.g.: five-day calendar]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskBound a useful workflow.# FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Definition of Done- Human-in-the-loop · review before decision# Workflow1. Define purpose and outcome.2. Bound start, finish, and risk.3. State decision and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not invent my context.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not execute.# Expected output- Workflow brief- Has scope and owner.- A case exists.- Boundary is written.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSWORKFLOW = from receiving the brief to delivering a reviewed proposalBOUNDARIES = starts with approved brief, ends with reviewed proposal and excludes negotiationRECORDS = five-day calendar with 17 blocks, 12 tasks and one delivered proposalADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Case: prepare and review a scope proposal. Start: brief received. End: version ready for a decision. Risk: citing an unverified source. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskBound a useful workflow.# FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Definition of Done- Human-in-the-loop · review before decision# Workflow1. Define purpose and outcome.2. Bound start, finish, and risk.3. State decision and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not invent my context.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not execute.# Expected output- Workflow brief- Has scope and owner.- A case exists.- Boundary is written.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When starting from a task or workflow that is not yet bounded.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSWORKFLOW = [WORKFLOW · work with a start, finish and output · e.g.: prepare a proposal]BOUNDARIES = [BOUNDARIES · workflow start, finish and exclusions · e.g.: from brief to approval]RECORDS = [RECORDS · observed calendar, task and delivery data · e.g.: five-day calendar]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Workflow briefInclude: Bound a useful workflow.Exclude: Does not execute.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Definition of Done- Human-in-the-loop · review before decisionSteps:1. Define purpose and outcome.2. Bound start, finish, and risk.3. State decision and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If no case exists, state that evidence is absent.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Workflow brief- Has scope and owner.- A case exists.- Boundary is written.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Has scope and owner.- A case exists.- Boundary is written.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Has scope and owner.; A case exists.; Boundary is written.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Less scope gives more control.## Provenance- Case data- Declared authorities: ev-m03-program, ev-m03-friction-playbook## Metadata- module: 03- prompt: prompt-01- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When starting from a task or workflow that is not yet bounded.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSWORKFLOW = from receiving the brief to delivering a reviewed proposalBOUNDARIES = starts with approved brief, ends with reviewed proposal and excludes negotiationRECORDS = five-day calendar with 17 blocks, 12 tasks and one delivered proposalADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Case: prepare and review a scope proposal. Start: brief received. End: version ready for a decision. Risk: citing an unverified source. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Workflow briefInclude: Bound a useful workflow.Exclude: Does not execute.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Definition of Done- Human-in-the-loop · review before decisionSteps:1. Define purpose and outcome.2. Bound start, finish, and risk.3. State decision and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If no case exists, state that evidence is absent.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Workflow brief- Has scope and owner.- A case exists.- Boundary is written.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Has scope and owner.- A case exists.- Boundary is written.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Has scope and owner.; A case exists.; Boundary is written.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Less scope gives more control.## Provenance- Synthetic demo- Declared authorities: ev-m03-program, ev-m03-friction-playbook## Metadata- module: 03- prompt: prompt-01- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Definition of Done- Human-in-the-loop · review before decision# Boundaries- Do not invent my context.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not execute.# Definition of DoneHas scope and owner.; A case exists.; Boundary is written.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSWORKFLOW = [WORKFLOW · work with a start, finish and output · e.g.: prepare a proposal]BOUNDARIES = [BOUNDARIES · workflow start, finish and exclusions · e.g.: from brief to approval]RECORDS = [RECORDS · observed calendar, task and delivery data · e.g.: five-day calendar]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskBound a useful workflow.# Workflow1. Define purpose and outcome.2. Bound start, finish, and risk.3. State decision and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Workflow brief- Has scope and owner.- A case exists.- Boundary is written.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Definition of Done- Human-in-the-loop · review before decision# Boundaries- Do not invent my context.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not execute.# Definition of DoneHas scope and owner.; A case exists.; Boundary is written.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSWORKFLOW = from receiving the brief to delivering a reviewed proposalBOUNDARIES = starts with approved brief, ends with reviewed proposal and excludes negotiationRECORDS = five-day calendar with 17 blocks, 12 tasks and one delivered proposalADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Case: prepare and review a scope proposal. Start: brief received. End: version ready for a decision. Risk: citing an unverified source. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskBound a useful workflow.# Workflow1. Define purpose and outcome.2. Bound start, finish, and risk.3. State decision and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Workflow brief- Has scope and owner.- A case exists.- Boundary is written.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Why it worksAcceptance criteria

Acceptance criteria

  • Has scope and owner.
  • A case exists.
  • Boundary is written.
  • The Mini-SOP retains a version, owner, exception and test evidence.
  • You can explain what you decided, the evidence used and what you will review next.

Edge cases

  • If no case exists, state that evidence is absent.
  • Workflow without a boundary: define start and finish before improving it.
  • Conflicting practices: record which one will be tested and why.

Trade-offs

  • Less scope gives more control.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • Does not execute.
LearnDiagnose frictionDiagnose friction.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you have the artifact from prompt 01.
Example
Workflow brief: brief → requirements → sources → draft → verification → decision. Final approval remains human. Owner and reviewer: you, at separate moments.
Evidence
Every friction has a moment.; Causes remain labeled.
Limit
Does not claim causality.
ReceivesWorkflow brief
ProducesFriction map
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <STEPS>current actions in their real orderExample: receive, analyze, draft, review · Required
  • <FRICTIONS>rework, waiting or late decisionsExample: incomplete citations and late review · Required
  • <EVIDENCE>facts or records supporting the reviewExample: board, citations and version · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Diagnose friction. The final decision and evidence remain under your responsibility.Data: STEPS=[STEPS · current actions in their real order · e.g.: receive, analyze, draft, review]; FRICTIONS=[FRICTIONS · rework, waiting or late decisions · e.g.: incomplete citations and late review]; EVIDENCE=[EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Friction map; Five Whys as questions; Human-in-the-loop.Order: Reconstruct the case. → Return the artifact, next step and any open gap.Deliver: Friction map; Every friction has a moment.Boundary: Do not assign causes.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Diagnose friction. The final decision and evidence remain under your responsibility.Data: STEPS=receive brief, find evidence, draft, review and deliver; FRICTIONS=two undated citations, one late review and 90 minutes of rework; EVIDENCE=five-day board, three citations and artifact version 0.2; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Workflow brief: brief → requirements → sources → draft → verification → decision. Final approval remains human. Owner and reviewer: you, at separate moments.Apply: Friction map; Five Whys as questions; Human-in-the-loop.Order: Reconstruct the case. → Return the artifact, next step and any open gap.Deliver: Friction map; Every friction has a moment.Boundary: Do not assign causes.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = [STEPS · current actions in their real order · e.g.: receive, analyze, draft, review]FRICTIONS = [FRICTIONS · rework, waiting or late decisions · e.g.: incomplete citations and late review]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDiagnose friction.# FRAMEWORKS AND BEST PRACTICES- Friction map- Five Whys as questions- Human-in-the-loop · review before decision# Workflow1. Reconstruct the case.2. Mark friction.3. Separate facts and hypotheses.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not assign causes.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not claim causality.# Expected output- Friction map- Every friction has a moment.- Causes remain labeled.- Observation is separate from hypothesis.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = receive brief, find evidence, draft, review and deliverFRICTIONS = two undated citations, one late review and 90 minutes of reworkEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Workflow brief: brief → requirements → sources → draft → verification → decision. Final approval remains human. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskDiagnose friction.# FRAMEWORKS AND BEST PRACTICES- Friction map- Five Whys as questions- Human-in-the-loop · review before decision# Workflow1. Reconstruct the case.2. Mark friction.3. Separate facts and hypotheses.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not assign causes.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not claim causality.# Expected output- Friction map- Every friction has a moment.- Causes remain labeled.- Observation is separate from hypothesis.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 01.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = [STEPS · current actions in their real order · e.g.: receive, analyze, draft, review]FRICTIONS = [FRICTIONS · rework, waiting or late decisions · e.g.: incomplete citations and late review]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Friction mapInclude: Diagnose friction.Exclude: Does not claim causality.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Friction map- Five Whys as questions- Human-in-the-loop · review before decisionSteps:1. Reconstruct the case.2. Mark friction.3. Separate facts and hypotheses.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If frictions compete, prioritize by impact and testability.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Friction map- Every friction has a moment.- Causes remain labeled.- Observation is separate from hypothesis.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Every friction has a moment.- Causes remain labeled.- Observation is separate from hypothesis.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Every friction has a moment.; Causes remain labeled.; Observation is separate from hypothesis.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Prudence reduces apparent certainty.## Provenance- Case data- Declared authorities: ev-m03-masterclass, ev-m03-friction-playbook## Metadata- module: 03- prompt: prompt-02- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 01.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = receive brief, find evidence, draft, review and deliverFRICTIONS = two undated citations, one late review and 90 minutes of reworkEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Workflow brief: brief → requirements → sources → draft → verification → decision. Final approval remains human. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Friction mapInclude: Diagnose friction.Exclude: Does not claim causality.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Friction map- Five Whys as questions- Human-in-the-loop · review before decisionSteps:1. Reconstruct the case.2. Mark friction.3. Separate facts and hypotheses.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If frictions compete, prioritize by impact and testability.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Friction map- Every friction has a moment.- Causes remain labeled.- Observation is separate from hypothesis.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Every friction has a moment.- Causes remain labeled.- Observation is separate from hypothesis.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Every friction has a moment.; Causes remain labeled.; Observation is separate from hypothesis.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Prudence reduces apparent certainty.## Provenance- Synthetic demo- Declared authorities: ev-m03-masterclass, ev-m03-friction-playbook## Metadata- module: 03- prompt: prompt-02- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Friction map- Five Whys as questions- Human-in-the-loop · review before decision# Boundaries- Do not assign causes.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not claim causality.# Definition of DoneEvery friction has a moment.; Causes remain labeled.; Observation is separate from hypothesis.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = [STEPS · current actions in their real order · e.g.: receive, analyze, draft, review]FRICTIONS = [FRICTIONS · rework, waiting or late decisions · e.g.: incomplete citations and late review]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDiagnose friction.# Workflow1. Reconstruct the case.2. Mark friction.3. Separate facts and hypotheses.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Friction map- Every friction has a moment.- Causes remain labeled.- Observation is separate from hypothesis.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Friction map- Five Whys as questions- Human-in-the-loop · review before decision# Boundaries- Do not assign causes.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not claim causality.# Definition of DoneEvery friction has a moment.; Causes remain labeled.; Observation is separate from hypothesis.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = receive brief, find evidence, draft, review and deliverFRICTIONS = two undated citations, one late review and 90 minutes of reworkEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Workflow brief: brief → requirements → sources → draft → verification → decision. Final approval remains human. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskDiagnose friction.# Workflow1. Reconstruct the case.2. Mark friction.3. Separate facts and hypotheses.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Friction map- Every friction has a moment.- Causes remain labeled.- Observation is separate from hypothesis.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Why it worksAcceptance criteria

Acceptance criteria

  • Every friction has a moment.
  • Causes remain labeled.
  • Observation is separate from hypothesis.
  • The Mini-SOP retains a version, owner, exception and test evidence.
  • You can explain what you decided, the evidence used and what you will review next.

Edge cases

  • If frictions compete, prioritize by impact and testability.
  • Workflow without a boundary: define start and finish before improving it.
  • Conflicting practices: record which one will be tested and why.

Trade-offs

  • Prudence reduces apparent certainty.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • Does not claim causality.
LearnFind relevant practicesFind relevant practices.4 · Prompt structure levelsTemplate / DemoOpen promptRun inFind sources · Deep Research
Before copying
When to use it
When you have the artifact from prompt 02.
Example
Friction map: waiting while finding sources; rework while fixing citations; repeated scope decisions; evidence: three recent reviews. Owner and reviewer: you, at separate moments.
Evidence
Retains provenance.; Explains differences.
Limit
Does not universalize recipes.
ReceivesFriction map
ProducesPractice dossier
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <QUESTION>testable question guiding the searchExample: what reduces review rework · Required
  • <SOURCES>materials with author, date and scopeExample: official guide and own record · Required
  • <ACCEPTANCE_CRITERION>observable condition for accepting the outputExample: citation, limit and recommendation · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Find relevant practices. The final decision and evidence remain under your responsibility.Data: QUESTION=[QUESTION · testable question guiding the search · e.g.: what reduces review rework]; SOURCES=[SOURCES · materials with author, date and scope · e.g.: official guide and own record]; ACCEPTANCE_CRITERION=[ACCEPTANCE_CRITERION · observable condition for accepting the output · e.g.: citation, limit and recommendation]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Evidence matrix; Context comparison; Human-in-the-loop.Order: Ask from friction. → Return the artifact, next step and any open gap.Deliver: Practice dossier; Retains provenance.Boundary: Use identifiable sources.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Find relevant practices. The final decision and evidence remain under your responsibility.Data: QUESTION=which controls reduce incomplete citations without delaying delivery; SOURCES=dated official guide and synthetic five-day record; ACCEPTANCE_CRITERION=each recommendation cites a source, states fit and proposes a test; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Friction map: waiting while finding sources; rework while fixing citations; repeated scope decisions; evidence: three recent reviews. Owner and reviewer: you, at separate moments.Apply: Evidence matrix; Context comparison; Human-in-the-loop.Order: Ask from friction. → Return the artifact, next step and any open gap.Deliver: Practice dossier; Retains provenance.Boundary: Use identifiable sources.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSQUESTION = [QUESTION · testable question guiding the search · e.g.: what reduces review rework]SOURCES = [SOURCES · materials with author, date and scope · e.g.: official guide and own record]ACCEPTANCE_CRITERION = [ACCEPTANCE_CRITERION · observable condition for accepting the output · e.g.: citation, limit and recommendation]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskFind relevant practices.# FRAMEWORKS AND BEST PRACTICES- Evidence matrix- Context comparison- Human-in-the-loop · review before decision# Workflow1. Ask from friction.2. Extract source and condition.3. Adapt, test, or reject.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Use identifiable sources.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not universalize recipes.# Expected output- Practice dossier- Retains provenance.- Explains differences.- Sources are accessible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSQUESTION = which controls reduce incomplete citations without delaying deliverySOURCES = dated official guide and synthetic five-day recordACCEPTANCE_CRITERION = each recommendation cites a source, states fit and proposes a testADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Friction map: waiting while finding sources; rework while fixing citations; repeated scope decisions; evidence: three recent reviews. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskFind relevant practices.# FRAMEWORKS AND BEST PRACTICES- Evidence matrix- Context comparison- Human-in-the-loop · review before decision# Workflow1. Ask from friction.2. Extract source and condition.3. Adapt, test, or reject.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Use identifiable sources.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not universalize recipes.# Expected output- Practice dossier- Retains provenance.- Explains differences.- Sources are accessible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 02.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSQUESTION = [QUESTION · testable question guiding the search · e.g.: what reduces review rework]SOURCES = [SOURCES · materials with author, date and scope · e.g.: official guide and own record]ACCEPTANCE_CRITERION = [ACCEPTANCE_CRITERION · observable condition for accepting the output · e.g.: citation, limit and recommendation]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Practice dossierInclude: Find relevant practices.Exclude: Does not universalize recipes.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Evidence matrix- Context comparison- Human-in-the-loop · review before decisionSteps:1. Ask from friction.2. Extract source and condition.3. Adapt, test, or reject.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If no case exists, state that evidence is absent.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Practice dossier- Retains provenance.- Explains differences.- Sources are accessible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Retains provenance.- Explains differences.- Sources are accessible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Retains provenance.; Explains differences.; Sources are accessible.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Traceability takes time.## Provenance- Case data- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-03- surface: source_searchReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 02.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSQUESTION = which controls reduce incomplete citations without delaying deliverySOURCES = dated official guide and synthetic five-day recordACCEPTANCE_CRITERION = each recommendation cites a source, states fit and proposes a testADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Friction map: waiting while finding sources; rework while fixing citations; repeated scope decisions; evidence: three recent reviews. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Practice dossierInclude: Find relevant practices.Exclude: Does not universalize recipes.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Evidence matrix- Context comparison- Human-in-the-loop · review before decisionSteps:1. Ask from friction.2. Extract source and condition.3. Adapt, test, or reject.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If no case exists, state that evidence is absent.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Practice dossier- Retains provenance.- Explains differences.- Sources are accessible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Retains provenance.- Explains differences.- Sources are accessible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Retains provenance.; Explains differences.; Sources are accessible.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Traceability takes time.## Provenance- Synthetic demo- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-03- surface: source_searchReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Evidence matrix- Context comparison- Human-in-the-loop · review before decision# Boundaries- Use identifiable sources.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not universalize recipes.# Definition of DoneRetains provenance.; Explains differences.; Sources are accessible.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSQUESTION = [QUESTION · testable question guiding the search · e.g.: what reduces review rework]SOURCES = [SOURCES · materials with author, date and scope · e.g.: official guide and own record]ACCEPTANCE_CRITERION = [ACCEPTANCE_CRITERION · observable condition for accepting the output · e.g.: citation, limit and recommendation]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskFind relevant practices.# Workflow1. Ask from friction.2. Extract source and condition.3. Adapt, test, or reject.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Practice dossier- Retains provenance.- Explains differences.- Sources are accessible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Evidence matrix- Context comparison- Human-in-the-loop · review before decision# Boundaries- Use identifiable sources.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not universalize recipes.# Definition of DoneRetains provenance.; Explains differences.; Sources are accessible.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSQUESTION = which controls reduce incomplete citations without delaying deliverySOURCES = dated official guide and synthetic five-day recordACCEPTANCE_CRITERION = each recommendation cites a source, states fit and proposes a testADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Friction map: waiting while finding sources; rework while fixing citations; repeated scope decisions; evidence: three recent reviews. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskFind relevant practices.# Workflow1. Ask from friction.2. Extract source and condition.3. Adapt, test, or reject.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Practice dossier- Retains provenance.- Explains differences.- Sources are accessible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Why it worksAcceptance criteria

Acceptance criteria

  • Retains provenance.
  • Explains differences.
  • Sources are accessible.
  • The Mini-SOP retains a version, owner, exception and test evidence.
  • You can explain what you decided, the evidence used and what you will review next.

Edge cases

  • If no case exists, state that evidence is absent.
  • Workflow without a boundary: define start and finish before improving it.
  • Conflicting practices: record which one will be tested and why.

Trade-offs

  • Traceability takes time.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • Does not universalize recipes.
LearnMap the current processMap the current process.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you have the artifact from prompt 03.
Example
Practice dossier: intake checklist, one source per claim and independent review. Limits: non-sensitive data and human approval. Owner and reviewer: you, at separate moments.
Evidence
Separates current and future.; Every decision has owner.
Limit
Does not design future.
ReceivesPractice dossier
ProducesCurrent process map
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <STEPS>current actions in their real orderExample: receive, analyze, draft, review · Required
  • <HANDOFFS>transfers between person, AI or teamExample: AI drafts; person validates · Required
  • <EXCEPTIONS>situations that justify breaking the ruleExample: client incident · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Map the current process. The final decision and evidence remain under your responsibility.Data: STEPS=[STEPS · current actions in their real order · e.g.: receive, analyze, draft, review]; HANDOFFS=[HANDOFFS · transfers between person, AI or team · e.g.: AI drafts; person validates]; EXCEPTIONS=[EXCEPTIONS · situations that justify breaking the rule · e.g.: client incident]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Light SIPOC; Minimal blueprint; Human-in-the-loop.Order: Order inputs and outputs. → Return the artifact, next step and any open gap.Deliver: Current process map; Separates current and future.Boundary: Do not complete missing facts.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Map the current process. The final decision and evidence remain under your responsibility.Data: STEPS=receive brief, find evidence, draft, review and deliver; HANDOFFS=AI proposes structure; analyst verifies citations; lead approves; EXCEPTIONS=client incident with immediate delivery impact; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Practice dossier: intake checklist, one source per claim and independent review. Limits: non-sensitive data and human approval. Owner and reviewer: you, at separate moments.Apply: Light SIPOC; Minimal blueprint; Human-in-the-loop.Order: Order inputs and outputs. → Return the artifact, next step and any open gap.Deliver: Current process map; Separates current and future.Boundary: Do not complete missing facts.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = [STEPS · current actions in their real order · e.g.: receive, analyze, draft, review]HANDOFFS = [HANDOFFS · transfers between person, AI or team · e.g.: AI drafts; person validates]EXCEPTIONS = [EXCEPTIONS · situations that justify breaking the rule · e.g.: client incident]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskMap the current process.# FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Minimal blueprint- Human-in-the-loop · review before decision# Workflow1. Order inputs and outputs.2. Assign decisions.3. Mark dependency and recovery.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not complete missing facts.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not design future.# Expected output- Current process map- Separates current and future.- Every decision has owner.- Decisions are explicit.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = receive brief, find evidence, draft, review and deliverHANDOFFS = AI proposes structure; analyst verifies citations; lead approvesEXCEPTIONS = client incident with immediate delivery impactADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Practice dossier: intake checklist, one source per claim and independent review. Limits: non-sensitive data and human approval. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskMap the current process.# FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Minimal blueprint- Human-in-the-loop · review before decision# Workflow1. Order inputs and outputs.2. Assign decisions.3. Mark dependency and recovery.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not complete missing facts.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not design future.# Expected output- Current process map- Separates current and future.- Every decision has owner.- Decisions are explicit.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 03.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = [STEPS · current actions in their real order · e.g.: receive, analyze, draft, review]HANDOFFS = [HANDOFFS · transfers between person, AI or team · e.g.: AI drafts; person validates]EXCEPTIONS = [EXCEPTIONS · situations that justify breaking the rule · e.g.: client incident]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Current process mapInclude: Map the current process.Exclude: Does not design future.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Minimal blueprint- Human-in-the-loop · review before decisionSteps:1. Order inputs and outputs.2. Assign decisions.3. Mark dependency and recovery.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If frictions compete, prioritize by impact and testability.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Current process map- Separates current and future.- Every decision has owner.- Decisions are explicit.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Separates current and future.- Every decision has owner.- Decisions are explicit.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Separates current and future.; Every decision has owner.; Decisions are explicit.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Detail may hide bottleneck.## Provenance- Case data- Declared authorities: ev-m03-friction-playbook, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-04- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 03.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = receive brief, find evidence, draft, review and deliverHANDOFFS = AI proposes structure; analyst verifies citations; lead approvesEXCEPTIONS = client incident with immediate delivery impactADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Practice dossier: intake checklist, one source per claim and independent review. Limits: non-sensitive data and human approval. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Current process mapInclude: Map the current process.Exclude: Does not design future.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Minimal blueprint- Human-in-the-loop · review before decisionSteps:1. Order inputs and outputs.2. Assign decisions.3. Mark dependency and recovery.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If frictions compete, prioritize by impact and testability.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Current process map- Separates current and future.- Every decision has owner.- Decisions are explicit.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Separates current and future.- Every decision has owner.- Decisions are explicit.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Separates current and future.; Every decision has owner.; Decisions are explicit.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Detail may hide bottleneck.## Provenance- Synthetic demo- Declared authorities: ev-m03-friction-playbook, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-04- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Minimal blueprint- Human-in-the-loop · review before decision# Boundaries- Do not complete missing facts.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not design future.# Definition of DoneSeparates current and future.; Every decision has owner.; Decisions are explicit.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = [STEPS · current actions in their real order · e.g.: receive, analyze, draft, review]HANDOFFS = [HANDOFFS · transfers between person, AI or team · e.g.: AI drafts; person validates]EXCEPTIONS = [EXCEPTIONS · situations that justify breaking the rule · e.g.: client incident]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskMap the current process.# Workflow1. Order inputs and outputs.2. Assign decisions.3. Mark dependency and recovery.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Current process map- Separates current and future.- Every decision has owner.- Decisions are explicit.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Light SIPOC- Minimal blueprint- Human-in-the-loop · review before decision# Boundaries- Do not complete missing facts.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not design future.# Definition of DoneSeparates current and future.; Every decision has owner.; Decisions are explicit.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSSTEPS = receive brief, find evidence, draft, review and deliverHANDOFFS = AI proposes structure; analyst verifies citations; lead approvesEXCEPTIONS = client incident with immediate delivery impactADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Practice dossier: intake checklist, one source per claim and independent review. Limits: non-sensitive data and human approval. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskMap the current process.# Workflow1. Order inputs and outputs.2. Assign decisions.3. Mark dependency and recovery.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Current process map- Separates current and future.- Every decision has owner.- Decisions are explicit.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Why it worksAcceptance criteria

Acceptance criteria

  • Separates current and future.
  • Every decision has owner.
  • Decisions are explicit.
  • The Mini-SOP retains a version, owner, exception and test evidence.
  • You can explain what you decided, the evidence used and what you will review next.

Edge cases

  • If frictions compete, prioritize by impact and testability.
  • Workflow without a boundary: define start and finish before improving it.
  • Conflicting practices: record which one will be tested and why.

Trade-offs

  • Detail may hide bottleneck.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • Does not design future.
EmbodyDefine the standardDefine the standard.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you have the artifact from prompt 04.
Example
Current process: request → brief → sources → AI-assisted draft → human verification → decision → version. Main friction: late sources. Owner and reviewer: you, at separate moments.
Evidence
Supports accept or reject.; Quality is not only speed.
Limit
Does not set priorities.
ReceivesCurrent process map
ProducesDefinition of done
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <OUTPUT>artifact that must be completeExample: proposal ready for decision · Required
  • <CRITERIA>explicit rules for comparing optionsExample: impact, urgency and effort · Required
  • <EVIDENCE>facts or records supporting the reviewExample: board, citations and version · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Define the standard. The final decision and evidence remain under your responsibility.Data: OUTPUT=[OUTPUT · artifact that must be complete · e.g.: proposal ready for decision]; CRITERIA=[CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]; EVIDENCE=[EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Definition of Done; Acceptance criteria; Human-in-the-loop.Order: Name outcome. → Return the artifact, next step and any open gap.Deliver: Definition of done; Supports accept or reject.Boundary: Do not prioritize for me.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Define the standard. The final decision and evidence remain under your responsibility.Data: OUTPUT=two-page proposal with citations and recommendation; CRITERIA=delivery impact, due date and remaining effort; EVIDENCE=five-day board, three citations and artifact version 0.2; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Current process: request → brief → sources → AI-assisted draft → human verification → decision → version. Main friction: late sources. Owner and reviewer: you, at separate moments.Apply: Definition of Done; Acceptance criteria; Human-in-the-loop.Order: Name outcome. → Return the artifact, next step and any open gap.Deliver: Definition of done; Supports accept or reject.Boundary: Do not prioritize for me.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSOUTPUT = [OUTPUT · artifact that must be complete · e.g.: proposal ready for decision]CRITERIA = [CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDefine the standard.# FRAMEWORKS AND BEST PRACTICES- Definition of Done- Acceptance criteria- Human-in-the-loop · review before decision# Workflow1. Name outcome.2. Write criteria.3. Add rejection and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not prioritize for me.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not set priorities.# Expected output- Definition of done- Supports accept or reject.- Quality is not only speed.- Criteria support rejection.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSOUTPUT = two-page proposal with citations and recommendationCRITERIA = delivery impact, due date and remaining effortEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Current process: request → brief → sources → AI-assisted draft → human verification → decision → version. Main friction: late sources. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskDefine the standard.# FRAMEWORKS AND BEST PRACTICES- Definition of Done- Acceptance criteria- Human-in-the-loop · review before decision# Workflow1. Name outcome.2. Write criteria.3. Add rejection and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not prioritize for me.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not set priorities.# Expected output- Definition of done- Supports accept or reject.- Quality is not only speed.- Criteria support rejection.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 04.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSOUTPUT = [OUTPUT · artifact that must be complete · e.g.: proposal ready for decision]CRITERIA = [CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Definition of doneInclude: Define the standard.Exclude: Does not set priorities.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Definition of Done- Acceptance criteria- Human-in-the-loop · review before decisionSteps:1. Name outcome.2. Write criteria.3. Add rejection and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If no case exists, state that evidence is absent.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Definition of done- Supports accept or reject.- Quality is not only speed.- Criteria support rejection.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Supports accept or reject.- Quality is not only speed.- Criteria support rejection.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Supports accept or reject.; Quality is not only speed.; Criteria support rejection.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: More rigor may be slower.## Provenance- Case data- Declared authorities: ev-m03-official-pdf, ev-m03-program, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-05- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 04.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSOUTPUT = two-page proposal with citations and recommendationCRITERIA = delivery impact, due date and remaining effortEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Current process: request → brief → sources → AI-assisted draft → human verification → decision → version. Main friction: late sources. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Definition of doneInclude: Define the standard.Exclude: Does not set priorities.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Definition of Done- Acceptance criteria- Human-in-the-loop · review before decisionSteps:1. Name outcome.2. Write criteria.3. Add rejection and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If no case exists, state that evidence is absent.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Definition of done- Supports accept or reject.- Quality is not only speed.- Criteria support rejection.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Supports accept or reject.- Quality is not only speed.- Criteria support rejection.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Supports accept or reject.; Quality is not only speed.; Criteria support rejection.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: More rigor may be slower.## Provenance- Synthetic demo- Declared authorities: ev-m03-official-pdf, ev-m03-program, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-05- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Definition of Done- Acceptance criteria- Human-in-the-loop · review before decision# Boundaries- Do not prioritize for me.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not set priorities.# Definition of DoneSupports accept or reject.; Quality is not only speed.; Criteria support rejection.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSOUTPUT = [OUTPUT · artifact that must be complete · e.g.: proposal ready for decision]CRITERIA = [CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDefine the standard.# Workflow1. Name outcome.2. Write criteria.3. Add rejection and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Definition of done- Supports accept or reject.- Quality is not only speed.- Criteria support rejection.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Definition of Done- Acceptance criteria- Human-in-the-loop · review before decision# Boundaries- Do not prioritize for me.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not set priorities.# Definition of DoneSupports accept or reject.; Quality is not only speed.; Criteria support rejection.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSOUTPUT = two-page proposal with citations and recommendationCRITERIA = delivery impact, due date and remaining effortEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Current process: request → brief → sources → AI-assisted draft → human verification → decision → version. Main friction: late sources. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskDefine the standard.# Workflow1. Name outcome.2. Write criteria.3. Add rejection and stop.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Definition of done- Supports accept or reject.- Quality is not only speed.- Criteria support rejection.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Why it worksAcceptance criteria

Acceptance criteria

  • Supports accept or reject.
  • Quality is not only speed.
  • Criteria support rejection.
  • The Mini-SOP retains a version, owner, exception and test evidence.
  • You can explain what you decided, the evidence used and what you will review next.

Edge cases

  • If no case exists, state that evidence is absent.
  • Workflow without a boundary: define start and finish before improving it.
  • Conflicting practices: record which one will be tested and why.

Trade-offs

  • More rigor may be slower.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • Does not set priorities.
EmbodyDraft the Mini-SOPDraft Mini-SOP v0.1.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you have the artifact from prompt 05.
Example
Completion criteria: scope covered, traceable sources, exceptions flagged and human approval recorded. Owner and reviewer: you, at separate moments.
Evidence
Can be followed without guessing.; Every exception has response.
Limit
Draft is not approval.
ReceivesDefinition of done
ProducesMini-SOP v0.1
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <PROCESS_MAP>current sequence with decisions and ownersExample: five-step map · Required
  • <CRITERIA>explicit rules for comparing optionsExample: impact, urgency and effort · Required
  • <EXCEPTIONS>situations that justify breaking the ruleExample: client incident · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Draft Mini-SOP v0.1. The final decision and evidence remain under your responsibility.Data: PROCESS_MAP=[PROCESS_MAP · current sequence with decisions and owners · e.g.: five-step map]; CRITERIA=[CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]; EXCEPTIONS=[EXCEPTIONS · situations that justify breaking the rule · e.g.: client incident]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Minimum SOP; Human-in-the-loop; Human-in-the-loop.Order: Integrate sequence. → Return the artifact, next step and any open gap.Deliver: Mini-SOP v0.1; Can be followed without guessing.Boundary: Do not approve the SOP.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Draft Mini-SOP v0.1. The final decision and evidence remain under your responsibility.Data: PROCESS_MAP=five-step map with two reviews and one approval point; CRITERIA=delivery impact, due date and remaining effort; EXCEPTIONS=client incident with immediate delivery impact; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Completion criteria: scope covered, traceable sources, exceptions flagged and human approval recorded. Owner and reviewer: you, at separate moments.Apply: Minimum SOP; Human-in-the-loop; Human-in-the-loop.Order: Integrate sequence. → Return the artifact, next step and any open gap.Deliver: Mini-SOP v0.1; Can be followed without guessing.Boundary: Do not approve the SOP.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSPROCESS_MAP = [PROCESS_MAP · current sequence with decisions and owners · e.g.: five-step map]CRITERIA = [CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]EXCEPTIONS = [EXCEPTIONS · situations that justify breaking the rule · e.g.: client incident]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDraft Mini-SOP v0.1.# FRAMEWORKS AND BEST PRACTICES- Minimum SOP- Human-in-the-loop- Human-in-the-loop · review before decision# Workflow1. Integrate sequence.2. Assign contributions.3. Add review and exception.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not approve the SOP.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Draft is not approval.# Expected output- Mini-SOP v0.1- Can be followed without guessing.- Every exception has response.- Review is placed.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSPROCESS_MAP = five-step map with two reviews and one approval pointCRITERIA = delivery impact, due date and remaining effortEXCEPTIONS = client incident with immediate delivery impactADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Completion criteria: scope covered, traceable sources, exceptions flagged and human approval recorded. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskDraft Mini-SOP v0.1.# FRAMEWORKS AND BEST PRACTICES- Minimum SOP- Human-in-the-loop- Human-in-the-loop · review before decision# Workflow1. Integrate sequence.2. Assign contributions.3. Add review and exception.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not approve the SOP.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Draft is not approval.# Expected output- Mini-SOP v0.1- Can be followed without guessing.- Every exception has response.- Review is placed.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 05.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSPROCESS_MAP = [PROCESS_MAP · current sequence with decisions and owners · e.g.: five-step map]CRITERIA = [CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]EXCEPTIONS = [EXCEPTIONS · situations that justify breaking the rule · e.g.: client incident]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Mini-SOP v0.1Include: Draft Mini-SOP v0.1.Exclude: Draft is not approval.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Minimum SOP- Human-in-the-loop- Human-in-the-loop · review before decisionSteps:1. Integrate sequence.2. Assign contributions.3. Add review and exception.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If frictions compete, prioritize by impact and testability.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Mini-SOP v0.1- Can be followed without guessing.- Every exception has response.- Review is placed.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Can be followed without guessing.- Every exception has response.- Review is placed.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Can be followed without guessing.; Every exception has response.; Review is placed.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Structure needs maintenance.## Provenance- Case data- Declared authorities: ev-m03-official-pdf, ev-m03-friction-playbook, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-06- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 05.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSPROCESS_MAP = five-step map with two reviews and one approval pointCRITERIA = delivery impact, due date and remaining effortEXCEPTIONS = client incident with immediate delivery impactADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Completion criteria: scope covered, traceable sources, exceptions flagged and human approval recorded. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Mini-SOP v0.1Include: Draft Mini-SOP v0.1.Exclude: Draft is not approval.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Minimum SOP- Human-in-the-loop- Human-in-the-loop · review before decisionSteps:1. Integrate sequence.2. Assign contributions.3. Add review and exception.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If frictions compete, prioritize by impact and testability.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Mini-SOP v0.1- Can be followed without guessing.- Every exception has response.- Review is placed.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Can be followed without guessing.- Every exception has response.- Review is placed.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Can be followed without guessing.; Every exception has response.; Review is placed.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Structure needs maintenance.## Provenance- Synthetic demo- Declared authorities: ev-m03-official-pdf, ev-m03-friction-playbook, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-06- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Minimum SOP- Human-in-the-loop- Human-in-the-loop · review before decision# Boundaries- Do not approve the SOP.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Draft is not approval.# Definition of DoneCan be followed without guessing.; Every exception has response.; Review is placed.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSPROCESS_MAP = [PROCESS_MAP · current sequence with decisions and owners · e.g.: five-step map]CRITERIA = [CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]EXCEPTIONS = [EXCEPTIONS · situations that justify breaking the rule · e.g.: client incident]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDraft Mini-SOP v0.1.# Workflow1. Integrate sequence.2. Assign contributions.3. Add review and exception.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Mini-SOP v0.1- Can be followed without guessing.- Every exception has response.- Review is placed.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Minimum SOP- Human-in-the-loop- Human-in-the-loop · review before decision# Boundaries- Do not approve the SOP.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Draft is not approval.# Definition of DoneCan be followed without guessing.; Every exception has response.; Review is placed.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSPROCESS_MAP = five-step map with two reviews and one approval pointCRITERIA = delivery impact, due date and remaining effortEXCEPTIONS = client incident with immediate delivery impactADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Completion criteria: scope covered, traceable sources, exceptions flagged and human approval recorded. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskDraft Mini-SOP v0.1.# Workflow1. Integrate sequence.2. Assign contributions.3. Add review and exception.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Mini-SOP v0.1- Can be followed without guessing.- Every exception has response.- Review is placed.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Why it worksAcceptance criteria

Acceptance criteria

  • Can be followed without guessing.
  • Every exception has response.
  • Review is placed.
  • The Mini-SOP retains a version, owner, exception and test evidence.
  • You can explain what you decided, the evidence used and what you will review next.

Edge cases

  • If frictions compete, prioritize by impact and testability.
  • Workflow without a boundary: define start and finish before improving it.
  • Conflicting practices: record which one will be tested and why.

Trade-offs

  • Structure needs maintenance.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • Draft is not approval.
EmbodyReview risks and exceptionsReview risks and exceptions.4 · Prompt structure levelsTemplate / DemoOpen promptRun inFind sources · Deep Research
Before copying
When to use it
When you have the artifact from prompt 06.
Example
Mini-SOP v0.1: collect brief; validate sources; prepare draft; review; record decision. Exception: stop when a source is missing. Owner and reviewer: you, at separate moments.
Evidence
Every risk has signal.; Decisions remain human.
Limit
Does not quantify without data.
ReceivesMini-SOP v0.1
ProducesRisk register
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <MINI_SOP>short versioned procedureExample: Mini-SOP v0.1 · Required
  • <RISKS>events that may derail the planExample: scope change · Required
  • <CONTROLS>checks required before proceedingExample: complete citations and approval · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Review risks and exceptions. The final decision and evidence remain under your responsibility.Data: MINI_SOP=[MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]; RISKS=[RISKS · events that may derail the plan · e.g.: scope change]; CONTROLS=[CONTROLS · checks required before proceeding · e.g.: complete citations and approval]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Pre-mortem; Risk register; Human-in-the-loop.Order: List failures. → Return the artifact, next step and any open gap.Deliver: Risk register; Every risk has signal.Boundary: Do not present risks as events.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Review risks and exceptions. The final decision and evidence remain under your responsibility.Data: MINI_SOP=Mini-SOP v0.1 for preparing sourced proposals; RISKS=scope change or delayed legal review; CONTROLS=dated citations, human review and block when authority is missing; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Mini-SOP v0.1: collect brief; validate sources; prepare draft; review; record decision. Exception: stop when a source is missing. Owner and reviewer: you, at separate moments.Apply: Pre-mortem; Risk register; Human-in-the-loop.Order: List failures. → Return the artifact, next step and any open gap.Deliver: Risk register; Every risk has signal.Boundary: Do not present risks as events.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = [MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]RISKS = [RISKS · events that may derail the plan · e.g.: scope change]CONTROLS = [CONTROLS · checks required before proceeding · e.g.: complete citations and approval]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskReview risks and exceptions.# FRAMEWORKS AND BEST PRACTICES- Pre-mortem- Risk register- Human-in-the-loop · review before decision# Workflow1. List failures.2. Assign prevention and response.3. Define gate and escalation.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not present risks as events.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not quantify without data.# Expected output- Risk register- Every risk has signal.- Decisions remain human.- Risks have responses.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = Mini-SOP v0.1 for preparing sourced proposalsRISKS = scope change or delayed legal reviewCONTROLS = dated citations, human review and block when authority is missingADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Mini-SOP v0.1: collect brief; validate sources; prepare draft; review; record decision. Exception: stop when a source is missing. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskReview risks and exceptions.# FRAMEWORKS AND BEST PRACTICES- Pre-mortem- Risk register- Human-in-the-loop · review before decision# Workflow1. List failures.2. Assign prevention and response.3. Define gate and escalation.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not present risks as events.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not quantify without data.# Expected output- Risk register- Every risk has signal.- Decisions remain human.- Risks have responses.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 06.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = [MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]RISKS = [RISKS · events that may derail the plan · e.g.: scope change]CONTROLS = [CONTROLS · checks required before proceeding · e.g.: complete citations and approval]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Risk registerInclude: Review risks and exceptions.Exclude: Does not quantify without data.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Pre-mortem- Risk register- Human-in-the-loop · review before decisionSteps:1. List failures.2. Assign prevention and response.3. Define gate and escalation.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If no case exists, state that evidence is absent.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Risk register- Every risk has signal.- Decisions remain human.- Risks have responses.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Every risk has signal.- Decisions remain human.- Risks have responses.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Every risk has signal.; Decisions remain human.; Risks have responses.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Controls slow but protect.## Provenance- Case data- Declared authorities: ev-m03-official-pdf, ev-m03-program, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-07- surface: source_searchReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 06.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = Mini-SOP v0.1 for preparing sourced proposalsRISKS = scope change or delayed legal reviewCONTROLS = dated citations, human review and block when authority is missingADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Mini-SOP v0.1: collect brief; validate sources; prepare draft; review; record decision. Exception: stop when a source is missing. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Risk registerInclude: Review risks and exceptions.Exclude: Does not quantify without data.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Pre-mortem- Risk register- Human-in-the-loop · review before decisionSteps:1. List failures.2. Assign prevention and response.3. Define gate and escalation.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If no case exists, state that evidence is absent.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Risk register- Every risk has signal.- Decisions remain human.- Risks have responses.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Every risk has signal.- Decisions remain human.- Risks have responses.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Every risk has signal.; Decisions remain human.; Risks have responses.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Controls slow but protect.## Provenance- Synthetic demo- Declared authorities: ev-m03-official-pdf, ev-m03-program, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-07- surface: source_searchReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Pre-mortem- Risk register- Human-in-the-loop · review before decision# Boundaries- Do not present risks as events.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not quantify without data.# Definition of DoneEvery risk has signal.; Decisions remain human.; Risks have responses.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = [MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]RISKS = [RISKS · events that may derail the plan · e.g.: scope change]CONTROLS = [CONTROLS · checks required before proceeding · e.g.: complete citations and approval]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskReview risks and exceptions.# Workflow1. List failures.2. Assign prevention and response.3. Define gate and escalation.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Risk register- Every risk has signal.- Decisions remain human.- Risks have responses.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Pre-mortem- Risk register- Human-in-the-loop · review before decision# Boundaries- Do not present risks as events.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not quantify without data.# Definition of DoneEvery risk has signal.; Decisions remain human.; Risks have responses.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = Mini-SOP v0.1 for preparing sourced proposalsRISKS = scope change or delayed legal reviewCONTROLS = dated citations, human review and block when authority is missingADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Mini-SOP v0.1: collect brief; validate sources; prepare draft; review; record decision. Exception: stop when a source is missing. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskReview risks and exceptions.# Workflow1. List failures.2. Assign prevention and response.3. Define gate and escalation.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Risk register- Every risk has signal.- Decisions remain human.- Risks have responses.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Why it worksAcceptance criteria

Acceptance criteria

  • Every risk has signal.
  • Decisions remain human.
  • Risks have responses.
  • The Mini-SOP retains a version, owner, exception and test evidence.
  • You can explain what you decided, the evidence used and what you will review next.

Edge cases

  • If no case exists, state that evidence is absent.
  • Workflow without a boundary: define start and finish before improving it.
  • Conflicting practices: record which one will be tested and why.

Trade-offs

  • Controls slow but protect.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • Does not quantify without data.
EmbodyDesign the testDesign a controlled test.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you have the artifact from prompt 07.
Example
Risk register: wrong citation/high/stop; sensitive data/high/do not upload; unclear scope/medium/request clarification. Owner and reviewer: you, at separate moments.
Evidence
Can stop safely.; Reconstructs the case.
Limit
Does not execute or approve.
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <MINI_SOP>short versioned procedureExample: Mini-SOP v0.1 · Required
  • <TEST_CASES>normal and adversarial casesExample: complete case and conflicting source · Required
  • <METRICS>measures used to compare the testExample: rework and cycle time · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Design a controlled test. The final decision and evidence remain under your responsibility.Data: MINI_SOP=[MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]; TEST_CASES=[TEST_CASES · normal and adversarial cases · e.g.: complete case and conflicting source]; METRICS=[METRICS · measures used to compare the test · e.g.: rework and cycle time]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: PDCA; PASA; Human-in-the-loop.Order: Set case and criteria. → Return the artifact, next step and any open gap.Deliver: Test plan; Can stop safely.Boundary: Do not execute the plan.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Design a controlled test. The final decision and evidence remain under your responsibility.Data: MINI_SOP=Mini-SOP v0.1 for preparing sourced proposals; TEST_CASES=normal case, conflicting source and incomplete brief; METRICS=rework minutes, cycle time and corrected citations; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Risk register: wrong citation/high/stop; sensitive data/high/do not upload; unclear scope/medium/request clarification. Owner and reviewer: you, at separate moments.Apply: PDCA; PASA; Human-in-the-loop.Order: Set case and criteria. → Return the artifact, next step and any open gap.Deliver: Test plan; Can stop safely.Boundary: Do not execute the plan.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = [MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]TEST_CASES = [TEST_CASES · normal and adversarial cases · e.g.: complete case and conflicting source]METRICS = [METRICS · measures used to compare the test · e.g.: rework and cycle time]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDesign a controlled test.# FRAMEWORKS AND BEST PRACTICES- PDCA- PASA- Human-in-the-loop · review before decision# Workflow1. Set case and criteria.2. Retain inputs.3. Define log and reviewer.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not execute the plan.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not execute or approve.# Expected output- Test plan- Can stop safely.- Reconstructs the case.- Case is reversible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = Mini-SOP v0.1 for preparing sourced proposalsTEST_CASES = normal case, conflicting source and incomplete briefMETRICS = rework minutes, cycle time and corrected citationsADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Risk register: wrong citation/high/stop; sensitive data/high/do not upload; unclear scope/medium/request clarification. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskDesign a controlled test.# FRAMEWORKS AND BEST PRACTICES- PDCA- PASA- Human-in-the-loop · review before decision# Workflow1. Set case and criteria.2. Retain inputs.3. Define log and reviewer.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not execute the plan.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not execute or approve.# Expected output- Test plan- Can stop safely.- Reconstructs the case.- Case is reversible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 07.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = [MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]TEST_CASES = [TEST_CASES · normal and adversarial cases · e.g.: complete case and conflicting source]METRICS = [METRICS · measures used to compare the test · e.g.: rework and cycle time]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Test planInclude: Design a controlled test.Exclude: Does not execute or approve.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- PDCA- PASA- Human-in-the-loop · review before decisionSteps:1. Set case and criteria.2. Retain inputs.3. Define log and reviewer.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If frictions compete, prioritize by impact and testability.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Test plan- Can stop safely.- Reconstructs the case.- Case is reversible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Can stop safely.- Reconstructs the case.- Case is reversible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Can stop safely.; Reconstructs the case.; Case is reversible.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: A test does not prove adoption.## Provenance- Case data- Declared authorities: ev-m03-official-pdf, ev-m03-program, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-08- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 07.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = Mini-SOP v0.1 for preparing sourced proposalsTEST_CASES = normal case, conflicting source and incomplete briefMETRICS = rework minutes, cycle time and corrected citationsADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Risk register: wrong citation/high/stop; sensitive data/high/do not upload; unclear scope/medium/request clarification. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Test planInclude: Design a controlled test.Exclude: Does not execute or approve.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- PDCA- PASA- Human-in-the-loop · review before decisionSteps:1. Set case and criteria.2. Retain inputs.3. Define log and reviewer.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If frictions compete, prioritize by impact and testability.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Test plan- Can stop safely.- Reconstructs the case.- Case is reversible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Can stop safely.- Reconstructs the case.- Case is reversible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Can stop safely.; Reconstructs the case.; Case is reversible.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: A test does not prove adoption.## Provenance- Synthetic demo- Declared authorities: ev-m03-official-pdf, ev-m03-program, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-08- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- PDCA- PASA- Human-in-the-loop · review before decision# Boundaries- Do not execute the plan.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not execute or approve.# Definition of DoneCan stop safely.; Reconstructs the case.; Case is reversible.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = [MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]TEST_CASES = [TEST_CASES · normal and adversarial cases · e.g.: complete case and conflicting source]METRICS = [METRICS · measures used to compare the test · e.g.: rework and cycle time]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDesign a controlled test.# Workflow1. Set case and criteria.2. Retain inputs.3. Define log and reviewer.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Test plan- Can stop safely.- Reconstructs the case.- Case is reversible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- PDCA- PASA- Human-in-the-loop · review before decision# Boundaries- Do not execute the plan.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not execute or approve.# Definition of DoneCan stop safely.; Reconstructs the case.; Case is reversible.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = Mini-SOP v0.1 for preparing sourced proposalsTEST_CASES = normal case, conflicting source and incomplete briefMETRICS = rework minutes, cycle time and corrected citationsADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Risk register: wrong citation/high/stop; sensitive data/high/do not upload; unclear scope/medium/request clarification. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskDesign a controlled test.# Workflow1. Set case and criteria.2. Retain inputs.3. Define log and reviewer.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Test plan- Can stop safely.- Reconstructs the case.- Case is reversible.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Why it worksAcceptance criteria

Acceptance criteria

  • Can stop safely.
  • Reconstructs the case.
  • Case is reversible.
  • The Mini-SOP retains a version, owner, exception and test evidence.
  • You can explain what you decided, the evidence used and what you will review next.

Edge cases

  • If frictions compete, prioritize by impact and testability.
  • Workflow without a boundary: define start and finish before improving it.
  • Conflicting practices: record which one will be tested and why.

Trade-offs

  • A test does not prove adoption.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • Does not execute or approve.
(R)EvolveAssess evidenceAssess evidence with AAR.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you have the artifact from prompt 08.
Example
Synthetic plan and log: one reversible case; 4 claims reviewed; 1 citation corrected; 0 sensitive data; human decision: adjust and repeat. Owner and reviewer: you, at separate moments.
Evidence
Conclusions cite records.; Changes receive decisions.
Limit
Does not invent evidence.
ReceivesTest plan + execution log
ProducesTest report
Consumes · Test plan
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <TEST_RECORD>observed results by caseExample: three documented runs · Required
  • <DEVIATIONS>differences from the procedureExample: skipped step or exception · Required
  • <VERDICT>decision to retain, adjust or stopExample: adjust before repeating · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Assess evidence with AAR. The final decision and evidence remain under your responsibility.Data: TEST_RECORD=[TEST_RECORD · observed results by case · e.g.: three documented runs]; DEVIATIONS=[DEVIATIONS · differences from the procedure · e.g.: skipped step or exception]; VERDICT=[VERDICT · decision to retain, adjust or stop · e.g.: adjust before repeating]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: AAR; Decision log; Human-in-the-loop.Order: Compare expectation and result. → Return the artifact, next step and any open gap.Deliver: Test report; Conclusions cite records.Boundary: Do not hide failures.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Assess evidence with AAR. The final decision and evidence remain under your responsibility.Data: TEST_RECORD=three runs: 42, 49 and 45 minutes; one corrected citation; DEVIATIONS=the date review was skipped in one of three runs; VERDICT=adjust the date control and repeat three cases; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Synthetic plan and log: one reversible case; 4 claims reviewed; 1 citation corrected; 0 sensitive data; human decision: adjust and repeat. Owner and reviewer: you, at separate moments.Apply: AAR; Decision log; Human-in-the-loop.Order: Compare expectation and result. → Return the artifact, next step and any open gap.Deliver: Test report; Conclusions cite records.Boundary: Do not hide failures.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSTEST_RECORD = [TEST_RECORD · observed results by case · e.g.: three documented runs]DEVIATIONS = [DEVIATIONS · differences from the procedure · e.g.: skipped step or exception]VERDICT = [VERDICT · decision to retain, adjust or stop · e.g.: adjust before repeating]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskAssess evidence with AAR.# FRAMEWORKS AND BEST PRACTICES- AAR- Decision log- Human-in-the-loop · review before decision# Workflow1. Compare expectation and result.2. Separate failure and exception.3. Submit changes to decision.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not hide failures.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not invent evidence.# Expected output- Test report- Conclusions cite records.- Changes receive decisions.- Conclusions link to log.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSTEST_RECORD = three runs: 42, 49 and 45 minutes; one corrected citationDEVIATIONS = the date review was skipped in one of three runsVERDICT = adjust the date control and repeat three casesADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Synthetic plan and log: one reversible case; 4 claims reviewed; 1 citation corrected; 0 sensitive data; human decision: adjust and repeat. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskAssess evidence with AAR.# FRAMEWORKS AND BEST PRACTICES- AAR- Decision log- Human-in-the-loop · review before decision# Workflow1. Compare expectation and result.2. Separate failure and exception.3. Submit changes to decision.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not hide failures.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not invent evidence.# Expected output- Test report- Conclusions cite records.- Changes receive decisions.- Conclusions link to log.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 08.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSTEST_RECORD = [TEST_RECORD · observed results by case · e.g.: three documented runs]DEVIATIONS = [DEVIATIONS · differences from the procedure · e.g.: skipped step or exception]VERDICT = [VERDICT · decision to retain, adjust or stop · e.g.: adjust before repeating]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Test reportInclude: Assess evidence with AAR.Exclude: Does not invent evidence.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- AAR- Decision log- Human-in-the-loop · review before decisionSteps:1. Compare expectation and result.2. Separate failure and exception.3. Submit changes to decision.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If no case exists, state that evidence is absent.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Test report- Conclusions cite records.- Changes receive decisions.- Conclusions link to log.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Conclusions cite records.- Changes receive decisions.- Conclusions link to log.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Conclusions cite records.; Changes receive decisions.; Conclusions link to log.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Honesty may stop scaling.## Provenance- Case data- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-09- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 08.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSTEST_RECORD = three runs: 42, 49 and 45 minutes; one corrected citationDEVIATIONS = the date review was skipped in one of three runsVERDICT = adjust the date control and repeat three casesADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Synthetic plan and log: one reversible case; 4 claims reviewed; 1 citation corrected; 0 sensitive data; human decision: adjust and repeat. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Test reportInclude: Assess evidence with AAR.Exclude: Does not invent evidence.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- AAR- Decision log- Human-in-the-loop · review before decisionSteps:1. Compare expectation and result.2. Separate failure and exception.3. Submit changes to decision.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If no case exists, state that evidence is absent.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Test report- Conclusions cite records.- Changes receive decisions.- Conclusions link to log.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Conclusions cite records.- Changes receive decisions.- Conclusions link to log.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Conclusions cite records.; Changes receive decisions.; Conclusions link to log.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Honesty may stop scaling.## Provenance- Synthetic demo- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-09- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- AAR- Decision log- Human-in-the-loop · review before decision# Boundaries- Do not hide failures.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not invent evidence.# Definition of DoneConclusions cite records.; Changes receive decisions.; Conclusions link to log.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSTEST_RECORD = [TEST_RECORD · observed results by case · e.g.: three documented runs]DEVIATIONS = [DEVIATIONS · differences from the procedure · e.g.: skipped step or exception]VERDICT = [VERDICT · decision to retain, adjust or stop · e.g.: adjust before repeating]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskAssess evidence with AAR.# Workflow1. Compare expectation and result.2. Separate failure and exception.3. Submit changes to decision.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Test report- Conclusions cite records.- Changes receive decisions.- Conclusions link to log.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- AAR- Decision log- Human-in-the-loop · review before decision# Boundaries- Do not hide failures.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not invent evidence.# Definition of DoneConclusions cite records.; Changes receive decisions.; Conclusions link to log.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSTEST_RECORD = three runs: 42, 49 and 45 minutes; one corrected citationDEVIATIONS = the date review was skipped in one of three runsVERDICT = adjust the date control and repeat three casesADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Synthetic plan and log: one reversible case; 4 claims reviewed; 1 citation corrected; 0 sensitive data; human decision: adjust and repeat. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskAssess evidence with AAR.# Workflow1. Compare expectation and result.2. Separate failure and exception.3. Submit changes to decision.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Test report- Conclusions cite records.- Changes receive decisions.- Conclusions link to log.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Why it worksAcceptance criteria

Acceptance criteria

  • Conclusions cite records.
  • Changes receive decisions.
  • Conclusions link to log.
  • The Mini-SOP retains a version, owner, exception and test evidence.
  • You can explain what you decided, the evidence used and what you will review next.

Edge cases

  • If no case exists, state that evidence is absent.
  • Workflow without a boundary: define start and finish before improving it.
  • Conflicting practices: record which one will be tested and why.

Trade-offs

  • Honesty may stop scaling.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • Does not invent evidence.
(R)EvolveVersion and transferVersion and transfer.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you have the artifact from prompt 09.
Example
Test report: 3 of 4 claims passed on the first attempt; one citation was corrected; decision: adopt with a checklist and second review. Owner and reviewer: you, at separate moments.
Evidence
Changes link evidence and owner.; Recipient can escalate.
Limit
Does not authorize deployment.
ReceivesTest report
ProducesMini-SOP v1.0
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <MINI_SOP>short versioned procedureExample: Mini-SOP v0.1 · Required
  • <CHANGES>justified changes for the new versionExample: add date control · Required
  • <EVIDENCE>facts or records supporting the reviewExample: board, citations and version · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Version and transfer. The final decision and evidence remain under your responsibility.Data: MINI_SOP=[MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]; CHANGES=[CHANGES · justified changes for the new version · e.g.: add date control]; EVIDENCE=[EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Version control; Transfer checklist; Human-in-the-loop.Order: Apply approved changes. → Return the artifact, next step and any open gap.Deliver: Mini-SOP v1.0; Changes link evidence and owner.Boundary: Do not generalize.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Version and transfer. The final decision and evidence remain under your responsibility.Data: MINI_SOP=Mini-SOP v0.1 for preparing sourced proposals; CHANGES=add date control and separate drafting from approval; EVIDENCE=five-day board, three citations and artifact version 0.2; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Test report: 3 of 4 claims passed on the first attempt; one citation was corrected; decision: adopt with a checklist and second review. Owner and reviewer: you, at separate moments.Apply: Version control; Transfer checklist; Human-in-the-loop.Order: Apply approved changes. → Return the artifact, next step and any open gap.Deliver: Mini-SOP v1.0; Changes link evidence and owner.Boundary: Do not generalize.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = [MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]CHANGES = [CHANGES · justified changes for the new version · e.g.: add date control]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskVersion and transfer.# FRAMEWORKS AND BEST PRACTICES- Version control- Transfer checklist- Human-in-the-loop · review before decision# Workflow1. Apply approved changes.2. State boundaries.3. Design second test.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not generalize.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not authorize deployment.# Expected output- Mini-SOP v1.0- Changes link evidence and owner.- Recipient can escalate.- Version states limits.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = Mini-SOP v0.1 for preparing sourced proposalsCHANGES = add date control and separate drafting from approvalEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Test report: 3 of 4 claims passed on the first attempt; one citation was corrected; decision: adopt with a checklist and second review. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskVersion and transfer.# FRAMEWORKS AND BEST PRACTICES- Version control- Transfer checklist- Human-in-the-loop · review before decision# Workflow1. Apply approved changes.2. State boundaries.3. Design second test.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Boundaries- Do not generalize.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not authorize deployment.# Expected output- Mini-SOP v1.0- Changes link evidence and owner.- Recipient can escalate.- Version states limits.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 09.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = [MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]CHANGES = [CHANGES · justified changes for the new version · e.g.: add date control]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Mini-SOP v1.0Include: Version and transfer.Exclude: Does not authorize deployment.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Version control- Transfer checklist- Human-in-the-loop · review before decisionSteps:1. Apply approved changes.2. State boundaries.3. Design second test.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If frictions compete, prioritize by impact and testability.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Mini-SOP v1.0- Changes link evidence and owner.- Recipient can escalate.- Version states limits.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Changes link evidence and owner.- Recipient can escalate.- Version states limits.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Changes link evidence and owner.; Recipient can escalate.; Version states limits.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Standardizing may freeze errors.## Provenance- Case data- Declared authorities: ev-m03-program, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-10- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you have the artifact from prompt 09.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = Mini-SOP v0.1 for preparing sourced proposalsCHANGES = add date control and separate drafting from approvalEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Test report: 3 of 4 claims passed on the first attempt; one citation was corrected; decision: adopt with a checklist and second review. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Mini-SOP v1.0Include: Version and transfer.Exclude: Does not authorize deployment.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Version control- Transfer checklist- Human-in-the-loop · review before decisionSteps:1. Apply approved changes.2. State boundaries.3. Design second test.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.### Edge cases- If frictions compete, prioritize by impact and testability.- Workflow without a boundary: define start and finish before improving it.- Conflicting practices: record which one will be tested and why.## C — Verifiable criteria# Expected output:- Mini-SOP v1.0- Changes link evidence and owner.- Recipient can escalate.- Version states limits.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.Observable criteria:- Changes link evidence and owner.- Recipient can escalate.- Version states limits.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.# Definition of Done: Changes link evidence and owner.; Recipient can escalate.; Version states limits.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.Trade-off: Standardizing may freeze errors.## Provenance- Synthetic demo- Declared authorities: ev-m03-program, ev-m03-ways-of-working## Metadata- module: 03- prompt: prompt-10- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Version control- Transfer checklist- Human-in-the-loop · review before decision# Boundaries- Do not generalize.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not authorize deployment.# Definition of DoneChanges link evidence and owner.; Recipient can escalate.; Version states limits.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = [MINI_SOP · short versioned procedure · e.g.: Mini-SOP v0.1]CHANGES = [CHANGES · justified changes for the new version · e.g.: add date control]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskVersion and transfer.# Workflow1. Apply approved changes.2. State boundaries.3. Design second test.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Mini-SOP v1.0- Changes link evidence and owner.- Recipient can escalate.- Version states limits.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Version control- Transfer checklist- Human-in-the-loop · review before decision# Boundaries- Do not generalize.- Keep the final decision human and request any missing input.- Do not claim improvement without comparing the workflow against an observed case.- Do not delegate a judgment that belongs to you to AI.- Does not authorize deployment.# Definition of DoneChanges link evidence and owner.; Recipient can escalate.; Version states limits.; The Mini-SOP retains a version, owner, exception and test evidence.; You can explain what you decided, the evidence used and what you will review next.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSMINI_SOP = Mini-SOP v0.1 for preparing sourced proposalsCHANGES = add date control and separate drafting from approvalEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Test report: 3 of 4 claims passed on the first attempt; one citation was corrected; decision: adopt with a checklist and second review. Owner and reviewer: you, at separate moments.# Optional adjustmentsState the missing fact that would change the decision.# TaskVersion and transfer.# Workflow1. Apply approved changes.2. State boundaries.3. Design second test.4. Check evidence, boundaries and consistency before closing.5. Return the artifact, next step and any open gap.# Expected output- Mini-SOP v1.0- Changes link evidence and owner.- Recipient can escalate.- Version states limits.- The Mini-SOP retains a version, owner, exception and test evidence.- You can explain what you decided, the evidence used and what you will review next.
Why it worksAcceptance criteria

Acceptance criteria

  • Changes link evidence and owner.
  • Recipient can escalate.
  • Version states limits.
  • The Mini-SOP retains a version, owner, exception and test evidence.
  • You can explain what you decided, the evidence used and what you will review next.

Edge cases

  • If frictions compete, prioritize by impact and testability.
  • Workflow without a boundary: define start and finish before improving it.
  • Conflicting practices: record which one will be tested and why.

Trade-offs

  • Standardizing may freeze errors.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • Does not authorize deployment.
For your practice

Turn one real task into a way of working you can repeat.

When every delivery depends on memory and improvisation, you repeat decisions, carry friction, and cannot tell what AI improved.

What you will gain

You will leave with a clear personal workflow, a test log, and a version you can reuse without losing control.

How to proceed

Choose a task, observe your practice, compare sources, co-create the Mini-SOP, and test one case under human review.

What to verify

A Mini-SOP with input, steps, criteria, human review, output, exceptions, and a real-case test log.

Design my workflow

Repeat the workflow on a second case and change only what the log supports.

4 metaprompts

Create reusable instruments to repeat and adapt the practice.

Capability ladder: from first practice to your own judgement.
Meta-promptCoach builderDesign a coach that teaches the module method from evidence and asks one question at a time.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you need to reuse an amplified, tested and transferable workflow without copying a previous session.
Example
Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.
Evidence
The instrument has an opening message, flow and close.; The rubric uses observable criteria.
Limit
It configures an instrument; it does not produce evidence by itself.
ReceivesDefinition of done
ProducesCoach instructions
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <COACH_OBJECTIVE>behavior the coach must teach and verifyExample: run an evidence-based review · Required
  • <AUDIENCE>person or team using the instrumentExample: professional with no prior context · Required
  • <SOURCES>materials with author, date and scopeExample: official guide and own record · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Design a coach that teaches the module method from evidence and asks one question at a time. The final decision and evidence remain under your responsibility.Data: COACH_OBJECTIVE=[COACH_OBJECTIVE · behavior the coach must teach and verify · e.g.: run an evidence-based review]; AUDIENCE=[AUDIENCE · person or team using the instrument · e.g.: professional with no prior context]; SOURCES=[SOURCES · materials with author, date and scope · e.g.: official guide and own record]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Source-grounded coaching; Teaching through questions; Adversarial test.Order: State purpose, audience and observable outcome. → Test the instrument with the Demo and record adjustments.Deliver: Coach instructions; The instrument has an opening message, flow and close.Boundary: Design the instrument; do not pretend it has been run.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Design a coach that teaches the module method from evidence and asks one question at a time. The final decision and evidence remain under your responsibility.Data: COACH_OBJECTIVE=guide an evidence-based review with one question at a time; AUDIENCE=professional preparing a first guided review; SOURCES=dated official guide and synthetic five-day record; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.Apply: Source-grounded coaching; Teaching through questions; Adversarial test.Order: State purpose, audience and observable outcome. → Test the instrument with the Demo and record adjustments.Deliver: Coach instructions; The instrument has an opening message, flow and close.Boundary: Design the instrument; do not pretend it has been run.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSCOACH_OBJECTIVE = [COACH_OBJECTIVE · behavior the coach must teach and verify · e.g.: run an evidence-based review]AUDIENCE = [AUDIENCE · person or team using the instrument · e.g.: professional with no prior context]SOURCES = [SOURCES · materials with author, date and scope · e.g.: official guide and own record]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDesign a coach that teaches the module method from evidence and asks one question at a time.# FRAMEWORKS AND BEST PRACTICES- Source-grounded coaching- Teaching through questions- Adversarial test# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Expected output- Coach instructions- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSCOACH_OBJECTIVE = guide an evidence-based review with one question at a timeAUDIENCE = professional preparing a first guided reviewSOURCES = dated official guide and synthetic five-day recordADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.# TaskDesign a coach that teaches the module method from evidence and asks one question at a time.# FRAMEWORKS AND BEST PRACTICES- Source-grounded coaching- Teaching through questions- Adversarial test# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Expected output- Coach instructions- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you need to reuse an amplified, tested and transferable workflow without copying a previous session.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSCOACH_OBJECTIVE = [COACH_OBJECTIVE · behavior the coach must teach and verify · e.g.: run an evidence-based review]AUDIENCE = [AUDIENCE · person or team using the instrument · e.g.: professional with no prior context]SOURCES = [SOURCES · materials with author, date and scope · e.g.: official guide and own record]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Coach instructionsInclude: Design a coach that teaches the module method from evidence and asks one question at a time.Exclude: It configures an instrument; it does not produce evidence by itself.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Source-grounded coaching- Teaching through questions- Adversarial testSteps:1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.### Edge cases- Evidence is missing: the instrument declares coverage_gap.- The request exceeds its role: it stops and states the boundary.- Ambiguous answer: it asks one concrete question before proceeding.## C — Verifiable criteria# Expected output:- Coach instructions- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.Observable criteria:- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.# Definition of Done: The instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.Trade-off: More structure improves repeatability but requires maintenance when sources change.## Provenance- Case data- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-masterclass## Metadata- module: 03- prompt: prompt-m1- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you need to reuse an amplified, tested and transferable workflow without copying a previous session.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSCOACH_OBJECTIVE = guide an evidence-based review with one question at a timeAUDIENCE = professional preparing a first guided reviewSOURCES = dated official guide and synthetic five-day recordADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Coach instructionsInclude: Design a coach that teaches the module method from evidence and asks one question at a time.Exclude: It configures an instrument; it does not produce evidence by itself.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Source-grounded coaching- Teaching through questions- Adversarial testSteps:1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.### Edge cases- Evidence is missing: the instrument declares coverage_gap.- The request exceeds its role: it stops and states the boundary.- Ambiguous answer: it asks one concrete question before proceeding.## C — Verifiable criteria# Expected output:- Coach instructions- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.Observable criteria:- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.# Definition of Done: The instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.Trade-off: More structure improves repeatability but requires maintenance when sources change.## Provenance- Synthetic demo- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-masterclass## Metadata- module: 03- prompt: prompt-m1- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Source-grounded coaching- Teaching through questions- Adversarial test# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Definition of DoneThe instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSCOACH_OBJECTIVE = [COACH_OBJECTIVE · behavior the coach must teach and verify · e.g.: run an evidence-based review]AUDIENCE = [AUDIENCE · person or team using the instrument · e.g.: professional with no prior context]SOURCES = [SOURCES · materials with author, date and scope · e.g.: official guide and own record]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDesign a coach that teaches the module method from evidence and asks one question at a time.# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Expected output- Coach instructions- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Source-grounded coaching- Teaching through questions- Adversarial test# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Definition of DoneThe instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSCOACH_OBJECTIVE = guide an evidence-based review with one question at a timeAUDIENCE = professional preparing a first guided reviewSOURCES = dated official guide and synthetic five-day recordADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.# TaskDesign a coach that teaches the module method from evidence and asks one question at a time.# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Expected output- Coach instructions- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Why it worksAcceptance criteria

Acceptance criteria

  • The instrument has an opening message, flow and close.
  • The rubric uses observable criteria.
  • You can explain what you decided, the evidence used and what you will review next.
  • The Mini-SOP retains a version, owner, exception and test evidence.

Edge cases

  • Evidence is missing: the instrument declares coverage_gap.
  • The request exceeds its role: it stops and states the boundary.
  • Ambiguous answer: it asks one concrete question before proceeding.

Trade-offs

  • More structure improves repeatability but requires maintenance when sources change.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • It configures an instrument; it does not produce evidence by itself.
Meta-promptEvaluator builderDesign a progressive assessment that checks application, judgment and transfer.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you need to reuse an amplified, tested and transferable workflow without copying a previous session.
Example
Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.
Evidence
The instrument has an opening message, flow and close.; The rubric uses observable criteria.
Limit
It configures an instrument; it does not produce evidence by itself.
ReceivesRisk register
ProducesAssessment plan
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <ASSESSMENT_GOAL>observable capability to verifyExample: apply the method to a case · Required
  • <EVIDENCE>facts or records supporting the reviewExample: board, citations and version · Required
  • <RUBRIC>observable criteria and levelsExample: apply, justify and transfer · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Design a progressive assessment that checks application, judgment and transfer. The final decision and evidence remain under your responsibility.Data: ASSESSMENT_GOAL=[ASSESSMENT_GOAL · observable capability to verify · e.g.: apply the method to a case]; EVIDENCE=[EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]; RUBRIC=[RUBRIC · observable criteria and levels · e.g.: apply, justify and transfer]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Observable rubric; Progressive assessment; Evidence-based retry.Order: State purpose, audience and observable outcome. → Test the instrument with the Demo and record adjustments.Deliver: Assessment plan; The instrument has an opening message, flow and close.Boundary: Design the instrument; do not pretend it has been run.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Design a progressive assessment that checks application, judgment and transfer. The final decision and evidence remain under your responsibility.Data: ASSESSMENT_GOAL=apply the method, justify the decision and transfer it; EVIDENCE=five-day board, three citations and artifact version 0.2; RUBRIC=three criteria with initial, sufficient and transferable levels; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.Apply: Observable rubric; Progressive assessment; Evidence-based retry.Order: State purpose, audience and observable outcome. → Test the instrument with the Demo and record adjustments.Deliver: Assessment plan; The instrument has an opening message, flow and close.Boundary: Design the instrument; do not pretend it has been run.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSASSESSMENT_GOAL = [ASSESSMENT_GOAL · observable capability to verify · e.g.: apply the method to a case]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]RUBRIC = [RUBRIC · observable criteria and levels · e.g.: apply, justify and transfer]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDesign a progressive assessment that checks application, judgment and transfer.# FRAMEWORKS AND BEST PRACTICES- Observable rubric- Progressive assessment- Evidence-based retry# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Expected output- Assessment plan- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSASSESSMENT_GOAL = apply the method, justify the decision and transfer itEVIDENCE = five-day board, three citations and artifact version 0.2RUBRIC = three criteria with initial, sufficient and transferable levelsADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.# TaskDesign a progressive assessment that checks application, judgment and transfer.# FRAMEWORKS AND BEST PRACTICES- Observable rubric- Progressive assessment- Evidence-based retry# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Expected output- Assessment plan- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you need to reuse an amplified, tested and transferable workflow without copying a previous session.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSASSESSMENT_GOAL = [ASSESSMENT_GOAL · observable capability to verify · e.g.: apply the method to a case]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]RUBRIC = [RUBRIC · observable criteria and levels · e.g.: apply, justify and transfer]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Assessment planInclude: Design a progressive assessment that checks application, judgment and transfer.Exclude: It configures an instrument; it does not produce evidence by itself.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Observable rubric- Progressive assessment- Evidence-based retrySteps:1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.### Edge cases- Evidence is missing: the instrument declares coverage_gap.- The request exceeds its role: it stops and states the boundary.- Ambiguous answer: it asks one concrete question before proceeding.## C — Verifiable criteria# Expected output:- Assessment plan- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.Observable criteria:- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.# Definition of Done: The instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.Trade-off: More structure improves repeatability but requires maintenance when sources change.## Provenance- Case data- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-masterclass## Metadata- module: 03- prompt: prompt-m2- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you need to reuse an amplified, tested and transferable workflow without copying a previous session.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSASSESSMENT_GOAL = apply the method, justify the decision and transfer itEVIDENCE = five-day board, three citations and artifact version 0.2RUBRIC = three criteria with initial, sufficient and transferable levelsADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Assessment planInclude: Design a progressive assessment that checks application, judgment and transfer.Exclude: It configures an instrument; it does not produce evidence by itself.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Observable rubric- Progressive assessment- Evidence-based retrySteps:1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.### Edge cases- Evidence is missing: the instrument declares coverage_gap.- The request exceeds its role: it stops and states the boundary.- Ambiguous answer: it asks one concrete question before proceeding.## C — Verifiable criteria# Expected output:- Assessment plan- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.Observable criteria:- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.# Definition of Done: The instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.Trade-off: More structure improves repeatability but requires maintenance when sources change.## Provenance- Synthetic demo- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-masterclass## Metadata- module: 03- prompt: prompt-m2- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Observable rubric- Progressive assessment- Evidence-based retry# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Definition of DoneThe instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSASSESSMENT_GOAL = [ASSESSMENT_GOAL · observable capability to verify · e.g.: apply the method to a case]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]RUBRIC = [RUBRIC · observable criteria and levels · e.g.: apply, justify and transfer]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDesign a progressive assessment that checks application, judgment and transfer.# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Expected output- Assessment plan- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Observable rubric- Progressive assessment- Evidence-based retry# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Definition of DoneThe instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSASSESSMENT_GOAL = apply the method, justify the decision and transfer itEVIDENCE = five-day board, three citations and artifact version 0.2RUBRIC = three criteria with initial, sufficient and transferable levelsADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.# TaskDesign a progressive assessment that checks application, judgment and transfer.# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Expected output- Assessment plan- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Why it worksAcceptance criteria

Acceptance criteria

  • The instrument has an opening message, flow and close.
  • The rubric uses observable criteria.
  • You can explain what you decided, the evidence used and what you will review next.
  • The Mini-SOP retains a version, owner, exception and test evidence.

Edge cases

  • Evidence is missing: the instrument declares coverage_gap.
  • The request exceeds its role: it stops and states the boundary.
  • Ambiguous answer: it asks one concrete question before proceeding.

Trade-offs

  • More structure improves repeatability but requires maintenance when sources change.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • It configures an instrument; it does not produce evidence by itself.
Meta-promptInterviewer builderDesign a demanding conversation rehearsal with objections tied to the module case.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you need to reuse an amplified, tested and transferable workflow without copying a previous session.
Example
Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.
Evidence
The instrument has an opening message, flow and close.; The rubric uses observable criteria.
Limit
It configures an instrument; it does not produce evidence by itself.
ReceivesTest report
ProducesInterview script
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <DELIVERABLE>document or conversation that must be readyExample: defensible review · Required
  • <OBJECTIONS>hard questions it must withstandExample: weak source or premature decision · Required
  • <EVIDENCE>facts or records supporting the reviewExample: board, citations and version · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Design a demanding conversation rehearsal with objections tied to the module case. The final decision and evidence remain under your responsibility.Data: DELIVERABLE=[DELIVERABLE · document or conversation that must be ready · e.g.: defensible review]; OBJECTIONS=[OBJECTIONS · hard questions it must withstand · e.g.: weak source or premature decision]; EVIDENCE=[EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: Objection map; Progressive difficulty; Verifiable feedback.Order: State purpose, audience and observable outcome. → Test the instrument with the Demo and record adjustments.Deliver: Interview script; The instrument has an opening message, flow and close.Boundary: Design the instrument; do not pretend it has been run.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Design a demanding conversation rehearsal with objections tied to the module case. The final decision and evidence remain under your responsibility.Data: DELIVERABLE=concise review with evidence, boundary and next step; OBJECTIONS=what evidence is missing and why this decision is reversible; EVIDENCE=five-day board, three citations and artifact version 0.2; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.Apply: Objection map; Progressive difficulty; Verifiable feedback.Order: State purpose, audience and observable outcome. → Test the instrument with the Demo and record adjustments.Deliver: Interview script; The instrument has an opening message, flow and close.Boundary: Design the instrument; do not pretend it has been run.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = [DELIVERABLE · document or conversation that must be ready · e.g.: defensible review]OBJECTIONS = [OBJECTIONS · hard questions it must withstand · e.g.: weak source or premature decision]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDesign a demanding conversation rehearsal with objections tied to the module case.# FRAMEWORKS AND BEST PRACTICES- Objection map- Progressive difficulty- Verifiable feedback# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Expected output- Interview script- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = concise review with evidence, boundary and next stepOBJECTIONS = what evidence is missing and why this decision is reversibleEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.# TaskDesign a demanding conversation rehearsal with objections tied to the module case.# FRAMEWORKS AND BEST PRACTICES- Objection map- Progressive difficulty- Verifiable feedback# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Expected output- Interview script- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you need to reuse an amplified, tested and transferable workflow without copying a previous session.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = [DELIVERABLE · document or conversation that must be ready · e.g.: defensible review]OBJECTIONS = [OBJECTIONS · hard questions it must withstand · e.g.: weak source or premature decision]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Interview scriptInclude: Design a demanding conversation rehearsal with objections tied to the module case.Exclude: It configures an instrument; it does not produce evidence by itself.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Objection map- Progressive difficulty- Verifiable feedbackSteps:1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.### Edge cases- Evidence is missing: the instrument declares coverage_gap.- The request exceeds its role: it stops and states the boundary.- Ambiguous answer: it asks one concrete question before proceeding.## C — Verifiable criteria# Expected output:- Interview script- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.Observable criteria:- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.# Definition of Done: The instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.Trade-off: More structure improves repeatability but requires maintenance when sources change.## Provenance- Case data- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-masterclass## Metadata- module: 03- prompt: prompt-m3- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you need to reuse an amplified, tested and transferable workflow without copying a previous session.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = concise review with evidence, boundary and next stepOBJECTIONS = what evidence is missing and why this decision is reversibleEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Interview scriptInclude: Design a demanding conversation rehearsal with objections tied to the module case.Exclude: It configures an instrument; it does not produce evidence by itself.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- Objection map- Progressive difficulty- Verifiable feedbackSteps:1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.### Edge cases- Evidence is missing: the instrument declares coverage_gap.- The request exceeds its role: it stops and states the boundary.- Ambiguous answer: it asks one concrete question before proceeding.## C — Verifiable criteria# Expected output:- Interview script- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.Observable criteria:- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.# Definition of Done: The instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.Trade-off: More structure improves repeatability but requires maintenance when sources change.## Provenance- Synthetic demo- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-masterclass## Metadata- module: 03- prompt: prompt-m3- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Objection map- Progressive difficulty- Verifiable feedback# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Definition of DoneThe instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = [DELIVERABLE · document or conversation that must be ready · e.g.: defensible review]OBJECTIONS = [OBJECTIONS · hard questions it must withstand · e.g.: weak source or premature decision]EVIDENCE = [EVIDENCE · facts or records supporting the review · e.g.: board, citations and version]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDesign a demanding conversation rehearsal with objections tied to the module case.# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Expected output- Interview script- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- Objection map- Progressive difficulty- Verifiable feedback# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Definition of DoneThe instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = concise review with evidence, boundary and next stepOBJECTIONS = what evidence is missing and why this decision is reversibleEVIDENCE = five-day board, three citations and artifact version 0.2ADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.# TaskDesign a demanding conversation rehearsal with objections tied to the module case.# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Expected output- Interview script- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Why it worksAcceptance criteria

Acceptance criteria

  • The instrument has an opening message, flow and close.
  • The rubric uses observable criteria.
  • You can explain what you decided, the evidence used and what you will review next.
  • The Mini-SOP retains a version, owner, exception and test evidence.

Edge cases

  • Evidence is missing: the instrument declares coverage_gap.
  • The request exceeds its role: it stops and states the boundary.
  • Ambiguous answer: it asks one concrete question before proceeding.

Trade-offs

  • More structure improves repeatability but requires maintenance when sources change.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • It configures an instrument; it does not produce evidence by itself.
Meta-promptPreparer builderDesign a guide that turns the module work into a concise, defensible deliverable.4 · Prompt structure levelsTemplate / DemoOpen promptRun inChat · selected sources
Before copying
When to use it
When you need to reuse an amplified, tested and transferable workflow without copying a previous session.
Example
Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.
Evidence
The instrument has an opening message, flow and close.; The rubric uses observable criteria.
Limit
It configures an instrument; it does not produce evidence by itself.
ReceivesMini-SOP v1.0
ProducesPreparation guide
Replace [FIELD IN CAPS] · adjust or delete [text in lower case]
What you need to replace 4
  • <DELIVERABLE>document or conversation that must be readyExample: defensible review · Required
  • <SECTIONS>minimum parts in orderExample: outcome, evidence and decision · Required
  • <CRITERIA>explicit rules for comparing optionsExample: impact, urgency and effort · Required
  • <ADOPTION_CRITERION>condition for adopting the workflow in your practiceExample: less rework with complete citations · Required
Parameters
  • LENGTH = concise
  • STRUCTURE = short table + decision
  • DEPTH = operational
  • APPROVAL = human required
Level 1 · DirectA clear instruction, ready to discuss and refine.
Level 1Direct

A clear instruction, ready to discuss and refine.

Objective: Design a guide that turns the module work into a concise, defensible deliverable. The final decision and evidence remain under your responsibility.Data: DELIVERABLE=[DELIVERABLE · document or conversation that must be ready · e.g.: defensible review]; SECTIONS=[SECTIONS · minimum parts in order · e.g.: outcome, evidence and decision]; CRITERIA=[CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]; ADOPTION_CRITERION=[ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations].Apply: BLUF; Definition of Done; Section rehearsal.Order: State purpose, audience and observable outcome. → Test the instrument with the Demo and record adjustments.Deliver: Preparation guide; The instrument has an opening message, flow and close.Boundary: Design the instrument; do not pretend it has been run.[State the missing fact that would change the decision.]
Demo · Level 1 · Direct
Objective: Design a guide that turns the module work into a concise, defensible deliverable. The final decision and evidence remain under your responsibility.Data: DELIVERABLE=concise review with evidence, boundary and next step; SECTIONS=outcome, evidence, decision, boundary and next step; CRITERIA=delivery impact, due date and remaining effort; ADOPTION_CRITERION=adopt if three cases reduce rework and retain complete citations; Available synthetic data: DEMO_ARTIFACT=Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.Apply: BLUF; Definition of Done; Section rehearsal.Order: State purpose, audience and observable outcome. → Test the instrument with the Demo and record adjustments.Deliver: Preparation guide; The instrument has an opening message, flow and close.Boundary: Design the instrument; do not pretend it has been run.State the missing fact that would change the decision.
Level 2 · StructuredMakes inputs, options, flow and expected result explicit.
Level 2Structured

Makes inputs, options, flow and expected result explicit.

# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = [DELIVERABLE · document or conversation that must be ready · e.g.: defensible review]SECTIONS = [SECTIONS · minimum parts in order · e.g.: outcome, evidence and decision]CRITERIA = [CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDesign a guide that turns the module work into a concise, defensible deliverable.# FRAMEWORKS AND BEST PRACTICES- BLUF- Definition of Done- Section rehearsal# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Expected output- Preparation guide- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Demo · Level 2 · Structured
# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = concise review with evidence, boundary and next stepSECTIONS = outcome, evidence, decision, boundary and next stepCRITERIA = delivery impact, due date and remaining effortADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.# TaskDesign a guide that turns the module work into a concise, defensible deliverable.# FRAMEWORKS AND BEST PRACTICES- BLUF- Definition of Done- Section rehearsal# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Expected output- Preparation guide- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Level 3 · SpecifiedAdds contract, boundaries, verification and definition of done.
Level 3Specified

Adds contract, boundaries, verification and definition of done.

# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you need to reuse an amplified, tested and transferable workflow without copying a previous session.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = [DELIVERABLE · document or conversation that must be ready · e.g.: defensible review]SECTIONS = [SECTIONS · minimum parts in order · e.g.: outcome, evidence and decision]CRITERIA = [CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Preparation guideInclude: Design a guide that turns the module work into a concise, defensible deliverable.Exclude: It configures an instrument; it does not produce evidence by itself.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- BLUF- Definition of Done- Section rehearsalSteps:1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.### Edge cases- Evidence is missing: the instrument declares coverage_gap.- The request exceeds its role: it stops and states the boundary.- Ambiguous answer: it asks one concrete question before proceeding.## C — Verifiable criteria# Expected output:- Preparation guide- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.Observable criteria:- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.# Definition of Done: The instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.Trade-off: More structure improves repeatability but requires maintenance when sources change.## Provenance- Case data- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-masterclass## Metadata- module: 03- prompt: prompt-m4- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Demo · Level 3 · Specified
# SPEC MetodologIAversion: 2.0status: executable## S — SituationContext: When you need to reuse an amplified, tested and transferable workflow without copying a previous session.# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = concise review with evidence, boundary and next stepSECTIONS = outcome, evidence, decision, boundary and next stepCRITERIA = delivery impact, due date and remaining effortADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.## P — Purpose and scopeExpert role: Act as a module specialist. Do not invent evidence or replace human judgment.Deliverable: Preparation guideInclude: Design a guide that turns the module work into a concise, defensible deliverable.Exclude: It configures an instrument; it does not produce evidence by itself.## E — Standards and execution### FRAMEWORKS AND BEST PRACTICES- BLUF- Definition of Done- Section rehearsalSteps:1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.### Edge cases- Evidence is missing: the instrument declares coverage_gap.- The request exceeds its role: it stops and states the boundary.- Ambiguous answer: it asks one concrete question before proceeding.## C — Verifiable criteria# Expected output:- Preparation guide- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.Observable criteria:- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.# Definition of Done: The instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.Trade-off: More structure improves repeatability but requires maintenance when sources change.## Provenance- Synthetic demo- Declared authorities: ev-m03-program, ev-m03-friction-playbook, ev-m03-masterclass## Metadata- module: 03- prompt: prompt-m4- surface: chatReturn conclusions, evidence and auditable metadata; do not expose private reasoning or chain of thought.
Level 4 · OrchestratedSeparates system instructions from the user request for reuse.
Level 4Orchestrated

Separates system instructions from the user request for reuse.

# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- BLUF- Definition of Done- Section rehearsal# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Definition of DoneThe instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = [DELIVERABLE · document or conversation that must be ready · e.g.: defensible review]SECTIONS = [SECTIONS · minimum parts in order · e.g.: outcome, evidence and decision]CRITERIA = [CRITERIA · explicit rules for comparing options · e.g.: impact, urgency and effort]ADOPTION_CRITERION = [ADOPTION_CRITERION · condition for adopting the workflow in your practice · e.g.: less rework with complete citations]# Optional adjustments[State the missing fact that would change the decision.]# TaskDesign a guide that turns the module work into a concise, defensible deliverable.# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Expected output- Preparation guide- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Demo · Level 4 · Orchestrated
# systemAct as a module specialist. Do not invent evidence or replace human judgment.# FRAMEWORKS AND BEST PRACTICES- BLUF- Definition of Done- Section rehearsal# Boundaries- Design the instrument; do not pretend it has been run.- Do not allow answers beyond the declared evidence.- Do not delegate a judgment that belongs to you to AI.- Do not claim improvement without comparing the workflow against an observed case.- It configures an instrument; it does not produce evidence by itself.# Definition of DoneThe instrument has an opening message, flow and close.; The rubric uses observable criteria.; You can explain what you decided, the evidence used and what you will review next.; The Mini-SOP retains a version, owner, exception and test evidence.# user# PARAMETERSLENGTH = conciseSTRUCTURE = short table + decisionDEPTH = operationalAPPROVAL = human required# INPUTSDELIVERABLE = concise review with evidence, boundary and next stepSECTIONS = outcome, evidence, decision, boundary and next stepCRITERIA = delivery impact, due date and remaining effortADOPTION_CRITERION = adopt if three cases reduce rework and retain complete citationsAvailable synthetic data: DEMO_ARTIFACT = Resolved synthetic case for your own practice: you retain the final decision and review the evidence; objective: an amplified, tested and transferable workflow; declared sources and boundaries.# Optional adjustmentsState the missing fact that would change the decision.# TaskDesign a guide that turns the module work into a concise, defensible deliverable.# Workflow1. State purpose, audience and observable outcome.2. Separate invariant rules from variable case data.3. Build the interaction flow and its gates.4. Add rubric, boundaries and adversarial cases.5. Test the instrument with the Demo and record adjustments.# Expected output- Preparation guide- The instrument has an opening message, flow and close.- The rubric uses observable criteria.- You can explain what you decided, the evidence used and what you will review next.- The Mini-SOP retains a version, owner, exception and test evidence.
Why it worksAcceptance criteria

Acceptance criteria

  • The instrument has an opening message, flow and close.
  • The rubric uses observable criteria.
  • You can explain what you decided, the evidence used and what you will review next.
  • The Mini-SOP retains a version, owner, exception and test evidence.

Edge cases

  • Evidence is missing: the instrument declares coverage_gap.
  • The request exceeds its role: it stops and states the boundary.
  • Ambiguous answer: it asks one concrete question before proceeding.

Trade-offs

  • More structure improves repeatability but requires maintenance when sources change.

Assumptions

  • The declared inputs, permissions and sources represent the case under review.

Limits

  • It configures an instrument; it does not produce evidence by itself.
Next step

Agentic work

Continue with the next resource in the journey.

Prompts → Agentic work