Imported from organvm-iv-taxis/a-i--skills (distributions/claude/skills/promotion-readiness-checklist/SKILL.md). Install upstream with npx skills add organvm-iv-taxis/a-i--skills --skill promotion-readiness-checklist. Copyright stays with the author (MIT).

Promotion Readiness Checklist

Assess whether a repository meets the requirements for promotion to the next governance state.

The Promotion State Machine

LOCAL → CANDIDATE → PUBLIC_PROCESS → GRADUATED → ARCHIVED
  ↓         ↓              ↓              ↓
 (no skip) (no skip)    (no skip)     (terminal)

Hard rule: No state skipping. Each transition requires explicit evidence of meeting all requirements.

LOCAL → CANDIDATE

Theme: "This exists and has basic hygiene."

Required

• Repository exists on GitHub with correct naming convention
• seed.yaml present with all required fields (schema_version, repo, organ, tier, promotion_status)
• README.md exists with ≥50 words explaining what this is
• .gitignore appropriate for the tech stack
• LICENSE file present (MIT for open, proprietary for internal)
• No secrets committed — checked with secret detection

Evidence

## LOCAL → CANDIDATE Evidence

- seed.yaml: ✓ (schema v1.0, organ: IV, tier: standard)
- README: ✓ (142 words)
- .gitignore: ✓ (Python template)
- LICENSE: ✓ (MIT)
- Secret scan: ✓ (clean)
- Date assessed: 2026-03-20
- Assessed by: {name/agent}

CANDIDATE → PUBLIC_PROCESS

Theme: "This is tested, documented, and CI-verified."

Required (all of LOCAL → CANDIDATE, plus)

• CI pipeline passing (lint + test on push/PR)
• Tests exist with ≥1 meaningful test
• CLAUDE.md or equivalent agent instructions present
• Code linting configured and passing (ruff, eslint, etc.)
• Dependency management configured (pyproject.toml, package.json)
• Branch protection enabled on main
• At least one successful deployment or release

Quality Metrics

Evidence

## CANDIDATE → PUBLIC_PROCESS Evidence

- CI: ✓ (GitHub Actions, 14/15 runs passed last 30 days = 93%)
- Tests: ✓ (12 tests, 45% coverage)
- CLAUDE.md: ✓ (includes dev commands, architecture, constraints)
- Linting: ✓ (ruff configured, 0 violations)
- Dependencies: ✓ (pyproject.toml with dev extras)
- Branch protection: ✓ (require PR, require CI)
- Deployment: ✓ (v0.1.0 released 2026-03-15)

PUBLIC_PROCESS → GRADUATED

Theme: "This is production-quality and externally comprehensible."

Required (all of CANDIDATE → PUBLIC_PROCESS, plus)

• Stranger test passed — documentation comprehensible to someone with zero context
• Architecture documentation with diagrams
• CONTRIBUTING.md with clear contribution guidelines
• Error handling comprehensive (no bare except: or swallowed errors)
• Security review completed (dependency audit, no known vulnerabilities)
• Performance baseline established (for applicable repos)
• Monitoring/observability in place (for applicable repos)
• Inter-repo dependencies documented in seed.yaml
• All downstream consumers identified and notified

Quality Metrics

Evidence

## PUBLIC_PROCESS → GRADUATED Evidence

- Stranger test: ✓ (Level 3 pass, 2026-03-18)
- Architecture: ✓ (docs/architecture.md with Mermaid diagrams)
- CONTRIBUTING: ✓ (includes setup, PR process, coding standards)
- Error handling: ✓ (reviewed, no bare excepts)
- Security: ✓ (no critical/high vulnerabilities, last scanned 2026-03-19)
- Performance: ✓ (P99 < 200ms, baseline established)
- Dependencies: ✓ (seed.yaml edges match actual imports)
- Coverage: 72%
- CI pass rate: 97%

GRADUATED → ARCHIVED

Theme: "This is no longer actively maintained."

Required

• Migration path documented for consumers
• Replacement identified (if applicable)
• Downstream consumers migrated or notified
• Final release tagged
• README updated with archived status and redirect
• GitHub repo archived (read-only)

Running the Assessment

Automated Assessment Script

def assess_promotion(repo_path: str, current_state: str, target_state: str) -> dict:
    checks = get_checks_for_transition(current_state, target_state)
    results = {}
    for check in checks:
        results[check.name] = check.evaluate(repo_path)

    passed = all(r["passed"] for r in results.values())
    return {
        "repo": repo_path,
        "transition": f"{current_state} → {target_state}",
        "passed": passed,
        "checks": results,
        "assessed_at": datetime.now().isoformat(),
    }

Promotion Request Template

## Promotion Request: {repo-name}

**Current state:** {state}
**Target state:** {state}
**Requested by:** {name}
**Date:** {date}

### Evidence
{Paste assessment output}

### Notes
{Any context about why this promotion is requested now}

Anti-Patterns

• Skipping states — The state machine exists for a reason; shortcuts create gaps
• Self-promotion without review — At least one other person/agent should verify
• Checking boxes without evidence — Each check needs concrete evidence, not just ✓
• Promoting to unblock — Promotion signals quality; don't dilute it for convenience
• No assessment record — Keep evidence of every promotion decision for audit
• One-time assessment — Graduated repos can regress; periodic re-assessment is healthy