/track-shadow-genome <context> <attempted_solution> <failure_mode>

- id: "SG-{sequential_number}"
  timestamp: "{ISO 8601 timestamp}"
  context: "{What we were building}"
  attempted_solution: "{What we tried}"
  failure_mode: "{From taxonomy above}"
  why_failed: "{Root cause analysis}"
  impact: "{Time lost, technical debt created, etc.}"
  lesson_learned: "{Actionable principle}"
  correct_approach: "{What worked instead}"
  related_entries: ["{Links to similar failures if applicable}"]
  preventability: "{Could this have been caught earlier? How?}"

## Shadow Genome Entry: SG-{number}

**Failure Mode:** {mode}
**Impact:** {impact}

**What We Tried:**
{attempted_solution}

**Why It Failed:**
{why_failed}

**Lesson Learned:**
{lesson_learned}

**Moving Forward:**
{correct_approach}

**Prevention:**
{How to avoid this in future}

## Shadow Genome Entry: SG-001

- id: "SG-001"
  timestamp: "2025-12-24T15:30:00Z"
  context: "Week 2 - Implementing database transaction safety"
  attempted_solution: "Use SQLAlchemy ORM for transaction management"
  failure_mode: "COMPLEXITY_VIOLATION"
  why_failed: "Added 5 new dependencies (50MB), introduced complexity in simple use case. Standard sqlite3 library has built-in transaction support."
  impact: "2 hours evaluating, 3 hours testing, 15MB production binary increase"
  lesson_learned: "Check stdlib first before adding dependencies. SQLite transactions are simple: conn.execute('BEGIN'), conn.commit(), conn.rollback()"
  correct_approach: "Manual transaction wrapper using stdlib sqlite3 - 10 lines of code, zero dependencies"
  preventability: "Could have been caught in architecture review with KISS checklist"

## Shadow Genome Entry: SG-002

- id: "SG-002"
  timestamp: "2025-12-26T10:00:00Z"
  context: "Week 3 - Validation dataset construction"
  attempted_solution: "Implement distributed processing with Celery for dataset generation"
  failure_mode: "PREMATURE_OPTIMIZATION"
  why_failed: "Dataset is 1000 examples, processes in 10 minutes single-threaded. Celery adds Redis dependency, deployment complexity. No measured bottleneck."
  impact: "1 day implementing Celery, 4 hours debugging Redis, added 200MB+ dependencies"
  lesson_learned: "Measure first, optimize second. 10 minutes is acceptable for weekly task. Only parallelize if >1 hour or run frequently."
  correct_approach: "Simple for-loop with tqdm progress bar. Fast enough, zero complexity."
  related_entries: ["SG-001"]
  preventability: "Pre-mortem would have identified: 'What if generation is fast enough without optimization?'"

## Shadow Genome Entry: SG-003

- id: "SG-003"
  timestamp: "2025-12-28T14:00:00Z"
  context: "Week 4 - Tier 3 formal verification design"
  attempted_solution: "Document that PyVeritas provides 100% verification coverage"
  failure_mode: "HALLUCINATION"
  why_failed: "PyVeritas research paper states ~80% accuracy. We claimed 100% without validation. Would have mislead users about system capabilities."
  impact: "Documentation would have been dishonest, violating Divergence Doctrine"
  lesson_learned: "ALWAYS cite exact numbers from source. Never round up. 80% ≠ 100%. Honest limitations build trust."
  correct_approach: "Document 'PyVeritas provides ~80% verification accuracy (per original research), complemented by Z3 for critical paths'"
  preventability: "Sentinel validation caught this before publication. Need to enforce citation accuracy checks."

# Q-DNA Development Shadow Genome
# Failed approaches archived for learning

metadata:
  version: "1.0"
  created: "2025-12-24"
  purpose: "Learn from failures, prevent repetition"

statistics:
  total_entries: 3
  most_common_failure: "COMPLEXITY_VIOLATION"
  prevention_rate: "67%" # (patterns detected / total entries)

failures:
  - id: "SG-001"
    # ... (as shown in examples)

  - id: "SG-002"
    # ... (as shown in examples)

patterns_detected:
  - pattern: "Premature dependency addition"
    occurrences: 2
    entries: ["SG-001", "SG-002"]
    prevention_mechanism: "Require measured bottleneck before new dependencies"
    status: "ACTIVE"

  - pattern: "Optimization before measurement"
    occurrences: 1
    entries: ["SG-002"]
    prevention_mechanism: "Benchmark-driven optimization only"
    status: "ACTIVE"

lessons_codified:
  - lesson: "Check stdlib before external dependencies"
    related_entries: ["SG-001"]
    enforced_by: "Architecture review checklist"

  - lesson: "Measure before optimizing"
    related_entries: ["SG-002"]
    enforced_by: "Performance testing required for optimization PRs"

  - lesson: "Cite exact numbers, not rounded approximations"
    related_entries: ["SG-003"]
    enforced_by: "Sentinel citation validation"

## Shadow Genome Updated

**Entry Added:** SG-004
**Failure Mode:** SECURITY_REGRESSION
**Impact:** Critical - keyfile integrity compromise
**Lesson:** All cryptographic changes require security review + external audit

**Pattern Detected:** 2 security regressions in 4 weeks
**Prevention Created:** Mandatory security review for all crypto PRs

**File Updated:** docs/SHADOW_GENOME.md
**Team Notification:** Posted to #development channel

**Action Required:**
- Update PR template to include security review checkbox
- Add crypto file watch in CI (auto-assign security reviewer)