ITIL Operational Implementation Guide
How to Integrate Framework into Daily PAI Operations
Version: 1.0
Created: 2026-06-21
Purpose: Step-by-step guide for operationalizing ITIL framework
Quick Start
1. Run Initial Metrics Collection
cd ~/.claude && bun PAI/Tools/ITILMetricsCollector.tsThis establishes your baseline across all 6 KPIs.
2. Assess Practice Maturity
bun PAI/Tools/PracticeMaturity.ts --weakest 5Identifies your 5 weakest practices for focused improvement.
3. Review Service Catalog
cat PAI/DOCUMENTATION/SERVICE_CATALOG.mdUnderstand what capabilities are available.
4. Set Up Monthly Review
Add to crontab or systemd timer:
# Run monthly review on 1st of each month
0 9 1 * * cd ~/.claude && bun PAI/Tools/ITILMonthlyReview.tsDaily Operations
Before Starting Work
1. Check applicable guiding principle:
| Task Type | Apply Principle |
|---|---|
| Bug fix / debugging | Start Where You Are - Reproduce before reading code |
| New feature | Focus on Value - Does this create principal value? |
| Refactoring | Keep It Simple - Remove complexity, don’t add |
| Multi-step work | Progress Iteratively - Break into phases, verify each |
| Cross-engine work | Collaborate & Visibility - Use shared ICM memory |
| Architecture decision | Think Systemically - What else does this affect? |
| Proven workflow | Optimize & Automate - Can this be a hook/skill? |
2. Run existence probe (Preflight Gate E):
# Before building anything new
rg "SEARCH_TERM" ~/.claude/PAI/ --type md -lEnsures you’re not duplicating existing work.
During Work
Algorithm OBSERVE Phase:
- Voice announcement: “Entering Algorithm”
- Auto-print applicable guiding principle (TODO: integrate into template)
- Run preflight gates
- Select capabilities
Algorithm REFLECT Phase:
- Auto-check: Did failed ISCs generate feedback memories?
- If no: Warning + prompt to create memory before phase=complete
After Work
Run Simplify:
# Post-implementation cleanup
bun PAI/Tools/Simplify.tsUpdate Metrics:
# Refresh ITIL dashboard
bun PAI/Tools/ITILMetricsCollector.tsWeekly Operations
Monday: Review Dashboard
Check ITIL CSF/KPI panel (once integrated):
- Are any KPIs red (fail)?
- What’s the trend (improving/declining)?
- Do any metrics need investigation?
Friday: Practice Check
Quick self-assessment:
- Which ITIL practice did I use most this week?
- Which practice did I skip that I should have used?
- Did I follow the seven guiding principles?
Monthly Operations
First Week of Month
1. Run Full Assessment:
cd ~/.claude
bun PAI/Tools/ITILMetricsCollector.ts
bun PAI/Tools/PracticeMaturity.ts
bun PAI/Tools/ITILMonthlyReview.ts2. Review Monthly Report:
Located at MEMORY/WORK/YYYYMM-01-000000_itil-monthly-review/REPORT.md
3. Implement Top Recommendation:
- Pick the highest-priority critical recommendation
- Create work item in TELOS or GitHub
- Schedule implementation this month
4. Update Service Catalog:
bun PAI/Tools/GenerateServiceCatalog.tsEnsures new skills/agents are documented.
Quarterly Operations
Review & Refine
1. TELOS Context Refresh: Use Interview skill to update principal context.
2. Practice Maturity Trends: Compare current maturity scores to 3 months ago:
# Review historical practice-maturity.json versions
git log -p -- MEMORY/STATE.claude/practice-maturity.json3. Framework Alignment Check:
- Are the seven principles still relevant?
- Do KPI targets need adjustment?
- Should any practices be added/removed?
4. Improvement Review:
- Which recommendations were implemented?
- What was the measurable impact?
- Which gaps remain?
Integration Points
Algorithm Template Updates
OBSERVE Phase (After Voice):
**📋 APPLICABLE GUIDING PRINCIPLE:**
[Auto-detect from task type and print relevant principle]
Examples:
- Bug fix → "Start Where You Are: Reproduce failure before code analysis"
- New feature → "Focus on Value: Every action must create stakeholder value"
- Refactor → "Keep It Simple: Use minimum steps to achieve objective"REFLECT Phase (Before phase=complete):
**✅ LEARNING CAPTURE CHECK:**
Failed ISCs: ${failedISCCount}
Feedback Memories Created: ${feedbackMemoryCount}
[If feedbackMemoryCount < failedISCCount]
⚠️ WARNING: ${failedISCCount - feedbackMemoryCount} failed ISCs lack feedback memories.
Create feedback memories before marking phase=complete.Hook Integration
New Hook: ITILPrincipleReminder.hook.ts
- Trigger: SessionStart
- Action: Print applicable guiding principle based on detected task type
- Implementation: Maps task keywords to principles
Enhanced Hook: PhaseTransitionGate.hook.ts
- Addition: Check feedback memory creation on REFLECT→COMPLETE transition
- Action: Block transition if failed ISCs have no memories (warn + bypass option)
Dashboard Panel
ITIL CSF/KPI Panel Structure:
┌─ ITIL Service Management ──────────────────┐
│ │
│ Value Delivery: 8.5/10 ✅ [▃▅▆▇█] │
│ Quality: 92% ⚠️ [▃▄▄▅▆] │
│ Learning Capture: 0.8 ❌ [▁▂▂▃▃] │
│ Efficiency: 45% ❌ [▁▁▂▂▃] │
│ Reliability: 99.5% ✅ [▆▇▇██] │
│ Security: 100% ✅ [█████] │
│ │
│ [View Details] [Monthly Review] │
└─────────────────────────────────────────────┘
Sparklines show 7-day trend, status symbols indicate vs. target.
Measurement Details
Data Collection Points
| Metric | Source | Collection Method |
|---|---|---|
| ISC Pass Rate | ISA files | Parse ## Criteria section, count [x] vs total |
| Euphoric Surprise | ISA frontmatter | Extract euphoric_surprise: field |
| Feedback Memories | MEMORY/Feedback/ | Count .md files |
| Ollama Routing | Inference.ts logs | Parse calls, calculate Ollama/(Ollama+Claude) % |
| Uptime | HealthCheck history | Track probe successes over time |
| Security Compliance | DestructiveOpGuard logs | Count blocks/(blocks+overrides) % |
Storage Format
Metrics: MEMORY/STATE.claude/itil-metrics.jsonl
{
"timestamp": "2026-06-21T12:00:00Z",
"valueDelivery": { "euphoricSurpriseAvg": 8.5, "target": 8, "status": "pass" },
"quality": { "iscPassRate": 92, "target": 95, "status": "warn" },
...
}Practice Maturity: MEMORY/STATE.claude/practice-maturity.json
{
"timestamp": "2026-06-21T12:00:00Z",
"practices": [
{ "practice": "Problem Detection", "score": 3, "evidence": [...], "gaps": [...] }
],
"avgScore": 2.8,
"weakest": ["Practice A", "Practice B", ...]
}Troubleshooting
”Metrics show 0% across the board”
Cause: No completed ISAs in MEMORY/WORK/ or paths incorrect.
Fix:
- Check PAI_DIR environment variable:
echo $PAI_DIR - Verify ISA files exist:
ls ~/.claude/PAI/MEMORY/WORK/*/ISA.md | head -5 - Re-run collector:
bun PAI/Tools/ITILMetricsCollector.ts - Inspect recent hook logs if the collector still reports zero metrics.
”Practice maturity scores too low”
Cause: Tool is checking ~/.claude paths, may be run from wrong environment.
Fix: Always run from PNC (Claude Code) environment where full PAI structure exists.
”Monthly review generates no recommendations”
Cause: All CSFs passing or insufficient metrics history.
Fix:
- Check metrics file:
wc -l MEMORY/STATE.claude/itil-metrics.jsonl - Need at least 2 metric snapshots for trends
- Run metrics collector multiple times over days
Success Criteria
After full implementation, you should see:
✅ Metrics collected automatically (weekly cron job)
✅ Dashboard shows all 6 KPIs with real-time status
✅ Failed ISCs auto-checked for feedback memories
✅ Guiding principles surface at decision points
✅ Monthly reviews generate actionable recommendations
✅ Practice maturity tracked with clear improvement path
✅ Service catalog up-to-date with all capabilities
Next Steps
- Immediate: Run initial baseline metrics
- This Week: Integrate guiding principles into Algorithm template
- This Month: Add ITIL panel to dashboard
- This Quarter: Implement top 3 practice maturity improvements
Related Files:
ITIL_FRAMEWORK.md- Complete framework documentationITIL_QUICK_REFERENCE.md- Daily-use cheat sheetSERVICE_CATALOG.md- Capability registryITILMetricsCollector.ts- Metrics collection toolPracticeMaturity.ts- Maturity assessment toolITILMonthlyReview.ts- Review automation
Questions? See framework documentation or run tools with --help flag.