persona-pillars-governance

Governance skill for OpenClaw core persona pillars (AGENTS.md, SOUL.md, IDENTITY.md, USER.md, TOOLS.md) with a scan → policy gate → guarded apply → snapshot rollback workflow.

What it does

• Scans five core context files and evaluates governance rules (scan.js, policy-engine.js)
• Produces actionable findings with PASS/REVIEW/BLOCK severity and evidence
• Applies managed-marker status updates only when gate is PASS (gated-apply.js)
• Creates snapshots before mutation and supports full restore (snapshot.js, restore.js)
• Generates unified reports and patch artifacts for audit/rollback traceability (report.js)

Safety model

• Managed write scope is marker-bounded (<!-- PPG:START ... --> ... <!-- PPG:END ... -->)
• readonly managed blocks are never auto-modified
• Gate result REVIEW/BLOCK blocks write path
• Snapshot + sha256 verification is required for restore consistency

Repository layout

.
├── SKILL.md
├── governance/
├── references/
├── scripts/persona-governance/
├── tests/persona-governance/
└── wrappers/

Quick start

# 1) scan
node scripts/persona-governance/scan.js --scope core --format json --root /root/.openclaw/workspace --strict

# 2) gated apply (uses latest scan/policy logs by default)
node scripts/persona-governance/gated-apply.js --run-id real-$(date -u +%Y%m%dT%H%M%SZ)

# 3) run tests
node scripts/persona-governance/run-tests.js

Recommended topics

• openclaw
• openclaw-skill
• persona-governance
• prompt-governance
• cron-automation
• nodejs

Release

Initial release: v1.0.0

• Policy-gated mutation path
• Marker-safe status autofix
• Snapshot/restore and traceable reporting
• Rulebook + rollback playbook included