danimerida2000/agents-governance-maintainer

Agents Governance Maintainer Skill

Mission

Keep the AI governance layer (AGENTS.md, docs/agent-context/**, .claude/skills/**) consistent with the actual state of the codebase after architecture, stack, route, schema, command, or policy changes.

Use When

• Architecture or stack changes have been made.
• New routes, schema tables, or commands have been added.
• Policy or governance rules need updating.
• Generated snapshots may be stale.
• Skills need updating after domain model changes.

Do Not Use

• For feature implementation (use the relevant engineering skill).
• For planning (use planner).
• For estimation (use feature-estimator).

Mandatory Inputs

1. AGENTS.md
2. docs/agent-context/README.md
3. All docs/agent-context/core/*.md files
4. All docs/agent-context/scoped/*.md files
5. .claude/skills/README.md
6. .claude/skills/edu-platform/SKILL.md
7. Root and workspace manifests (package.json, Cargo.toml)
8. Live source files in packages/**
9. Scripts in scripts/docs/** and scripts/git/**
10. Docker and infrastructure files

Workflow

1. Identify the scope of drift (policy, architecture, stack, routes, schema, commands, tooling).
2. Inspect live manifests and source files before trusting any generated snapshot.
3. Update the affected governance files.
4. Regenerate generated snapshots if applicable.
5. Run validation checks.
6. Update the active work document.
7. Record evidence in the execution log.

Required Output

1. Updated governance files with changes documented.
2. Generated snapshots refreshed (if applicable).
3. Validation checks passing.
4. Work document updated with evidence.

Validation

node scripts/docs/check-agent-links.mjs
node scripts/docs/generate-doc-inventories.mjs --check
bash scripts/git/check-quality-gate.sh

References

• Codebase Surfaces
• Sync Checklist