Process Standards¶

Purpose: Artifact requirements, naming conventions, and executor interview standards for the OODA workflow
Status: ACTIVE
Version: 1.0.0
Last Updated: 2025-12-19

📋 Required Artifacts by Phase¶

Purpose: Explicit list of MANDATORY artifacts for each workflow phase to ensure completeness and prevent progression gaps.

CAPTURE Phase Artifacts¶

REQUIRED (Must Create - Artifacts)¶

✅ idea.md - Capability idea and original prompt

REQUIRED (Must Create - Validation)¶

✅ capture-validation.md - Quality validation (@quality-assurance, score ≥6/10)
✅ capture-to-observe-gate.md - Gate decision (@gate-keeper, PASS/FAIL)

REQUIRED (Must Create - Memory)¶

✅ step-1.md through step-6.md - Step memories (150-250 tokens each)
✅ capture-phase-summary.md - Phase summary (≤250 tokens)

OBSERVE Phase Artifacts¶

REQUIRED (Must Create - Artifacts)¶

✅ initial-jtbd-frame.md - Initial Jobs-to-be-Done framing
✅ interviews/executors/*.md - Minimum 3 executor interviews (see Executor Interview Requirements)
✅ interviews/beneficiaries/*.md - 3-6 beneficiary interviews
✅ job-definition.md - Tool-agnostic job definition
✅ job-journey.md - Current state journey with tools
✅ outcome-gaps.md - Gaps between desired and actual outcomes

REQUIRED (Must Create - Validation)¶

✅ observe-validation.md - Quality validation (@quality-assurance, score ≥8/10)
✅ observe-advisory-review.md - Advisory alignment review (score ≥8/10)
✅ observe-to-orient-gate.md - Gate decision (@gate-keeper, PASS/FAIL)

REQUIRED (Must Create - Memory)¶

✅ step-1.md through step-8.md - Step memories (150-250 tokens each)
✅ observe-phase-summary.md - Phase summary (≤900 tokens)

ORIENT Phase Artifacts¶

REQUIRED (Must Create - Artifacts)¶

✅ symptoms-analysis.md - Observable symptoms and pain points
✅ diagnosis.md - Root cause analysis
✅ prognosis.md - Future state projections if unaddressed
✅ problem-statement.md - Structured problem definition
✅ kano-classification.md - Problem categorization (must-be/attractive/indifferent)
✅ rice-prioritization.md - RICE scoring for problems
✅ problem-prioritization.md - Final prioritized problem list

REQUIRED (Must Create - Validation)¶

✅ orient-validation.md - Quality validation (@quality-assurance, score ≥8/10)
✅ orient-advisory-review.md - Advisory alignment review (score ≥8/10)
✅ orient-to-recommend-gate.md - Gate decision (@gate-keeper, PASS/FAIL)

REQUIRED (Must Create - Memory)¶

✅ step-1.md through step-11.md - Step memories (150-250 tokens each)
✅ orient-phase-summary.md - Phase summary (≤900 tokens)

REQUIRED (Must Create - Artifacts)¶

✅ solution-hypotheses.md - Initial solution ideas (5-10 hypotheses)
✅ blue-ocean-analysis.md - Blue Ocean Strategy canvas
✅ solution-rice.md - RICE scoring for solutions
✅ solution-prioritization.md - Final prioritized solution list
✅ recommendation-selection.md - Selected solution with rationale
✅ recommendation.md - Toulmin-method recommendation

REQUIRED (Must Create - Validation)¶

✅ recommend-validation.md - Quality validation (@quality-assurance, score ≥8/10)
✅ recommend-advisory-review.md - Advisory alignment review (score ≥8/10)
✅ recommend-to-align-gate.md - Gate decision (@gate-keeper, PASS/FAIL)

REQUIRED (Must Create - Memory)¶

✅ step-1.md through step-8.md - Step memories (150-250 tokens each)
✅ recommend-phase-summary.md - Phase summary (≤900 tokens)

ALIGN Phase Artifacts¶

REQUIRED (Must Create - Artifacts)¶

✅ capability-thesis.md - Complete capability proposal
✅ execution-plan.md - RGRD-based implementation plan
✅ alignment-record.md - Decision record and commitments

REQUIRED (Must Create - Validation)¶

✅ align-to-ready-gate.md - Gate decision (@gate-keeper, PASS/FAIL)

REQUIRED (Must Create - Memory)¶

✅ step-1.md through step-4.md - Step memories (150-250 tokens each)
✅ align-phase-summary.md - Phase summary (≤900 tokens)
✅ workflow-summary.md - Complete workflow summary (root level, 3,000-4,000 tokens)

CONDITIONAL¶

⚠️ disagreements-log.md - Only if disagreements occurred during alignment
⚠️ align-validation.md - Optional quality validation

👥 Executor Interview Requirements¶

OBSERVE Phase - Executor Interview Requirements¶

REQUIRED (Minimum): Create 3 core executor interviews covering primary perspectives:

Core Interviews (Mandatory)¶

Blue Hat (Strategic/Innovator): Forward-thinking, process-oriented, innovation-focused
White Hat (Data-Driven/Analytical): Facts-focused, evidence-based, objective analysis
Green Hat (Creative/Problem-Solver): Creative thinking, alternatives generation

RECOMMENDED (Additional): Add 3 supplementary interviews for full perspective diversity:

Supplementary Interviews (Recommended for Complex/High-Risk Capabilities)¶

Yellow Hat (Optimist/Benefits-Focused): Opportunity identification, positive outcomes
Black Hat (Skeptic/Risk-Aware): Risk analysis, critical evaluation, constraints
Red Hat (Pragmatist/Emotion-Driven): Intuitive feedback, emotional response, gut checks

Decision Criteria for Supplementary Interviews: - ✅ Include Black Hat if: High technical risk, regulatory concerns, or failure impact > 3 weeks - ✅ Include Yellow Hat if: New market opportunity, unproven value hypothesis, innovation focus - ✅ Include Red Hat if: User-facing capability, emotional/UX component, change management needed

Validation Rules: - Minimum 3 core interviews MUST be present - If complexity score ≥ 7/10, SHOULD include Black Hat - Document rationale if omitting recommended interviews

File Naming: - interviews/executors/blue-hat-innovator.md (required) - interviews/executors/white-hat-early-adopter.md (required) - interviews/executors/green-hat-early-majority.md (required) - interviews/executors/yellow-hat-optimist.md (optional) - interviews/executors/black-hat-skeptic.md (optional) - interviews/executors/red-hat-pragmatist.md (optional)

🏷️ Artifact File Naming Standards (STRICT)¶

Purpose: Ensure consistency, enable automation, prevent validation failures.

Rule: All artifact filenames listed below are MANDATORY and EXACT. No variations, abbreviations, or substitutions allowed.

Enforcement: Validation gates check for exact filename matches.

Complete Artifact Name Registry¶

CAPTURE Phase: - ✅ idea.md (EXACT)

OBSERVE Phase: - ✅ initial-jtbd-frame.md (EXACT) - ✅ interviews/executors/<hat>-<persona>.md (e.g., blue-hat-innovator.md) - ✅ interviews/beneficiaries/<name>.md (e.g., enterprise-admin.md) - ✅ job-definition.md (EXACT) - ✅ job-journey.md (EXACT) - ✅ outcome-gaps.md (EXACT) - ✅ observe-validation.md (EXACT) - ⚠️ observe-advisory-review.md (EXACT, optional)

ORIENT Phase: - ✅ symptoms-analysis.md (EXACT) - ✅ diagnosis.md (EXACT) - ✅ prognosis.md (EXACT) - ✅ problem-statement.md (EXACT) - ✅ kano-classification.md (EXACT) - ✅ rice-prioritization.md (EXACT) - ✅ problem-prioritization.md (EXACT) - ✅ orient-validation.md (EXACT)

RECOMMEND Phase: - ✅ solution-hypotheses.md (EXACT) - ✅ blue-ocean-analysis.md (EXACT)
- ❌ WRONG: blue-ocean-canvas.md, blue-ocean-strategy.md, bos-analysis.md - ✅ solution-rice.md (EXACT)
- ❌ WRONG: solution-rice-scores.md, rice-solutions.md, solution-scoring.md - ✅ solution-prioritization.md (EXACT) - ✅ recommendation-selection.md (EXACT)
- ❌ WRONG: selected-recommendation.md, recommendation-choice.md - ✅ recommendation.md (EXACT) - ✅ recommend-validation.md (EXACT)

ALIGN Phase: - ✅ capability-thesis.md (EXACT) - ✅ execution-plan.md (EXACT) - ✅ alignment-record.md (EXACT) - ⚠️ disagreements-log.md (EXACT, conditional)

Memory Files: - ✅ <phase>/memory/step-<N>.md (e.g., step-1.md, step-2.md)
- ❌ WRONG: step-01.md, <phase>-step-1.md, step-3-4.md - ✅ <phase>/memory/<phase>-phase-summary.md (e.g., observe-phase-summary.md) - ✅ workflow-summary.md (EXACT, root level)

Validation Files: - ✅ <phase>/validation/<phase>-validation.md (e.g., observe-validation.md) - ✅ <phase>/validation/<phase>-advisory-review.md (e.g., observe-advisory-review.md) - ✅ <phase>/validation/<from-phase>-to-<to-phase>-gate.md (e.g., observe-to-orient-gate.md)

Version Tracking: - ✅ version.json (EXACT)

Naming Convention Examples¶

Correct Examples:

✅ blue-ocean-analysis.md
✅ solution-rice.md
✅ recommendation-selection.md
✅ recommend/memory/step-1.md
✅ recommend/memory/recommend-phase-summary.md

Incorrect Examples:

❌ blue-ocean-canvas.md (use blue-ocean-analysis.md)
❌ solution-rice-scores.md (use solution-rice.md)
❌ selected-recommendation.md (use recommendation-selection.md)
❌ .memory/recommend-step-1.md (use recommend/memory/step-1.md)
❌ recommend/memory/step-01.md (use recommend/memory/step-1.md)

Rationale for Exact Naming¶

Automation: Scripts and tools rely on predictable filenames
Validation: Gate checks use exact filename matching
Documentation: Templates reference specific filenames
Consistency: Reduces cognitive load, improves cross-version consistency
Debugging: Easier to identify missing or misnamed files

Template Alignment¶

Each artifact name corresponds to a template in .contributing/templates/: - solution-rice.md → .contributing/templates/solution-rice.md.template - blue-ocean-analysis.md → .contributing/templates/blue-ocean-analysis.md.template - recommendation-selection.md → .contributing/templates/recommendation-selection.md.template - step-1.md → .contributing/templates/step-memory.md.template - observe-validation.md → .contributing/templates/phase-validation.md.template - observe-advisory-review.md → .contributing/templates/phase-advisory-review.md.template - observe-to-orient-gate.md → .contributing/templates/gate-decision.md.template

Rule: Template filename (minus .template) = Artifact filename (or category for reusable templates)

🔄 Workflow Completeness Requirements¶

Step Memory Requirements¶

MANDATORY: Every workflow step MUST create a step memory.

Step Count by Phase¶

CAPTURE: 6 steps → 6 step memories
OBSERVE: 8 steps → 8 step memories
ORIENT: 11 steps → 11 step memories (includes Kano+RICE split and conditional revision)
RECOMMEND: 8 steps → 8 step memories
ALIGN: 4 steps → 4 step memories (includes conditional disagreements step)
TOTAL: 37 step memories maximum (35 typical, 2 conditional)

Step Memory Naming¶

Format: <phase>/memory/step-N.md (e.g., step-1.md, step-2.md)

Examples: - observe/memory/step-1.md (Initial JTBD Frame) - observe/memory/step-2.md (Executor Interviews) - orient/memory/step-5.md (Kano Classification) - orient/memory/step-6.md (RICE Prioritization) - orient/memory/step-7.md (Combined Prioritization) - orient/memory/step-8.md (Revision if quality < 8/10 - may contain "Skipped: score ≥8/10")

Step Memory Template¶

See .contributing/templates/step-memory.md.template

Required Elements: - Step number and total (e.g., "4/8 (OBSERVE)") - Agent name (@analyst, @ux-researcher, etc.) - Framework used (if applicable) - Completion timestamp (ISO 8601) - What was done (1-2 sentences) - Key outputs (3-5 bullets) - Key insights (2-3 bullets) - Next step (or "Phase complete")

Token Budget: 150-250 tokens per step memory

Validation Artifact Requirements¶

MANDATORY: Every phase MUST create quality validation and gate decisions. Advisory reviews required for OBSERVE, ORIENT, RECOMMEND.

Validation Artifacts by Phase¶

Phase	Quality Validation	Advisory Review	Gate Decision
CAPTURE	✅ capture-validation.md	❌ N/A	✅ capture-to-observe-gate.md
OBSERVE	✅ observe-validation.md	✅ observe-advisory-review.md	✅ observe-to-orient-gate.md
ORIENT	✅ orient-validation.md	✅ orient-advisory-review.md	✅ orient-to-recommend-gate.md
RECOMMEND	✅ recommend-validation.md	✅ recommend-advisory-review.md	✅ recommend-to-align-gate.md
ALIGN	⚠️ align-validation.md (optional)	❌ N/A	✅ align-to-ready-gate.md

Total Required: 4-5 quality validations, 3 advisory reviews, 5 gate decisions (12-13 validation artifacts)

Validation Artifact Locations¶

<phase>/validation/<phase>-validation.md
<phase>/validation/<phase>-advisory-review.md
<phase>/validation/<from-phase>-to-<to-phase>-gate.md

Templates¶

.contributing/templates/phase-validation.md.template
.contributing/templates/phase-advisory-review.md.template
.contributing/templates/gate-decision.md.template

Validation Scoring Requirements¶

Quality Validation: Must score ≥8/10 to proceed to gate
Advisory Review: Must score ≥8/10 alignment to proceed to gate
Gate Decision: Must PASS all criteria to unlock next phase

Phase Summary Requirements¶

MANDATORY: Every phase MUST create a phase summary before transition.

Phase Summary Specifications¶

Phase	Summary Path	Token Budget	Key Content
CAPTURE	capture/memory/capture-phase-summary.md	≤250	Capability name, prompt (≤50 words), context, validation score
OBSERVE	observe/memory/observe-phase-summary.md	≤900	Job definition, executors, critical findings, top 5 gaps
ORIENT	orient/memory/orient-phase-summary.md	≤900	Problem (medical model), prioritization (RICE+Kano), scope
RECOMMEND	recommend/memory/recommend-phase-summary.md	≤900	Solution (Toulmin), MVP components, success metrics, alternatives
ALIGN	align/memory/align-phase-summary.md	≤900	Capability thesis, commitment, execution plan, stakeholder alignment

Workflow Summary: workflow-summary.md (root level, 3,000-4,000 tokens, created in ALIGN Step 4)

Phase Summary Template¶

See .contributing/templates/phase-memory.md.template

Required Elements: - Phase name and completion timestamp - Quality score (X/10) - Advisory score (X.X/10, if applicable) - Gate status (PASSED/FAILED) - Key findings (3-5 bullets) - Critical artifacts created - Next phase guidance

Completeness Enforcement¶

Pre-Flight Checks (Before Phase Start)¶

Before starting a new phase, verify the prior phase is complete:

# Verify prior phase summary exists
if [ ! -f "$CAPABILITY_PATH/$PRIOR_PHASE/memory/$PRIOR_PHASE-phase-summary.md" ]; then
  echo "❌ Error: Cannot start $CURRENT_PHASE - $PRIOR_PHASE phase summary missing"
  exit 1
fi

# Verify gate passed
if ! grep -q "Decision: PASS" "$CAPABILITY_PATH/$PRIOR_PHASE/validation/*-gate.md"; then
  echo "❌ Error: Cannot start $CURRENT_PHASE - $PRIOR_PHASE gate failed"
  exit 1
fi

Post-Flight Checks (After Phase Complete)¶

After completing a phase, verify all required artifacts:

# Verify step memories (adjust EXPECTED_STEPS per phase)
EXPECTED_STEPS=8  # OBSERVE: 8, ORIENT: 11, RECOMMEND: 8, ALIGN: 4
ACTUAL_STEPS=$(ls $CAPABILITY_PATH/$PHASE/memory/step-*.md 2>/dev/null | wc -l)

if [ "$ACTUAL_STEPS" -ne "$EXPECTED_STEPS" ]; then
  echo "⚠️  Warning: Step memory gap - Found $ACTUAL_STEPS, expected $EXPECTED_STEPS"
fi

# Verify validation artifacts
if [ ! -f "$CAPABILITY_PATH/$PHASE/validation/$PHASE-validation.md" ]; then
  echo "❌ Error: Missing $PHASE-validation.md"
  exit 1
fi

# Verify phase summary
if [ ! -f "$CAPABILITY_PATH/$PHASE/memory/$PHASE-phase-summary.md" ]; then
  echo "❌ Error: Missing $PHASE-phase-summary.md"
  exit 1
fi

# Verify gate decision
if [ ! -f "$CAPABILITY_PATH/$PHASE/validation/"*"-gate.md" ]; then
  echo "❌ Error: Missing gate decision"
  exit 1
fi

Token Budget Enforcement¶

# Check phase summary token budget
SUMMARY_FILE="$CAPABILITY_PATH/$PHASE/memory/$PHASE-phase-summary.md"
TOKEN_COUNT=$(wc -w < "$SUMMARY_FILE")  # Approximate: 1 token ≈ 0.75 words

MAX_TOKENS=900  # 250 for CAPTURE, 900 for others
if [ "$TOKEN_COUNT" -gt "$MAX_TOKENS" ]; then
  echo "⚠️  Warning: $PHASE phase summary exceeds token budget ($TOKEN_COUNT > $MAX_TOKENS)"
  echo "    Please compress to meet budget"
fi

Session-Per-Phase Architecture¶

RECOMMENDED: Execute each OODA phase in a separate AI session to maximize context retention and reduce token usage.

Session Benefits¶

Token Efficiency: 95% reduction (75,000 → 3,500 tokens per phase)
Context Retention: Fresh context for each phase vs compounded decay
Error Isolation: Phase failures don't contaminate subsequent phases
Parallel Execution: Multiple phases can be explored simultaneously

Session Pattern¶

Session Start: 1. Load prior phase summary (~250-900 tokens) 2. Load workflow guide for current phase 3. Execute phase steps 4. Create all step memories (150-250 tokens each) 5. Create phase summary (≤250 or ≤900 tokens) 6. Run validation and gate checks

Session End: - Commit phase summary to memory - Document gate decision - Hand off to next session via phase summary

Example (OBSERVE → ORIENT Transition):

OBSERVE Session End:
- Creates observe-phase-summary.md (≤900 tokens)
- Creates observe-to-orient-gate.md (PASS)
- Exits

ORIENT Session Start:
- Loads observe-phase-summary.md (~900 tokens, not 25,000+)
- Executes ORIENT workflow
- Creates orient-phase-summary.md (≤900 tokens)
- Exits

Related Documentation: - AGENTS.md - Main agent navigation - MEMORY-PROTOCOL.md - Memory management rules - VALIDATION-GATES.md - Phase transition gates

Referenced By: - .contributing/agent/artifact-validator.md - Artifact validation agent - .contributing/agent/naming-enforcer.md - Naming validation agent - All workflow phase agents

Process Standards¶

📋 Required Artifacts by Phase¶

CAPTURE Phase Artifacts¶

REQUIRED (Must Create - Artifacts)¶

REQUIRED (Must Create - Validation)¶

REQUIRED (Must Create - Memory)¶

OBSERVE Phase Artifacts¶

REQUIRED (Must Create - Artifacts)¶

REQUIRED (Must Create - Validation)¶

REQUIRED (Must Create - Memory)¶

ORIENT Phase Artifacts¶

REQUIRED (Must Create - Artifacts)¶

REQUIRED (Must Create - Validation)¶

REQUIRED (Must Create - Memory)¶

RECOMMEND Phase Artifacts¶

REQUIRED (Must Create - Artifacts)¶

REQUIRED (Must Create - Validation)¶

REQUIRED (Must Create - Memory)¶

ALIGN Phase Artifacts¶

REQUIRED (Must Create - Artifacts)¶

REQUIRED (Must Create - Validation)¶

REQUIRED (Must Create - Memory)¶

CONDITIONAL¶

👥 Executor Interview Requirements¶

OBSERVE Phase - Executor Interview Requirements¶

Core Interviews (Mandatory)¶

Supplementary Interviews (Recommended for Complex/High-Risk Capabilities)¶

🏷️ Artifact File Naming Standards (STRICT)¶

Complete Artifact Name Registry¶

Naming Convention Examples¶

Rationale for Exact Naming¶

Template Alignment¶

🔄 Workflow Completeness Requirements¶

Step Memory Requirements¶

Step Count by Phase¶

Step Memory Naming¶

Step Memory Template¶

Validation Artifact Requirements¶

Validation Artifacts by Phase¶

Validation Artifact Locations¶

Templates¶

Validation Scoring Requirements¶

Phase Summary Requirements¶

Phase Summary Specifications¶

Phase Summary Template¶

Completeness Enforcement¶

Pre-Flight Checks (Before Phase Start)¶

Post-Flight Checks (After Phase Complete)¶

Token Budget Enforcement¶

Session-Per-Phase Architecture¶

Session Benefits¶

Session Pattern¶

Navigation¶