config-guardian — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited config-guardian (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Rule #1: Read the manual. Rule #2: Propose diff, do not apply. Rule #3: Log everything.
One skill. Three modes. Universal.
Before touching ANY config, verify knowledge is current:
# Adapt to the system
openclaw status # OpenClaw
docker --version # Docker
psql --version # PostgreSQL
arifos health # arifOSUse web_fetch or local docs (/app/docs/ for OpenClaw).
State explicitly:
| Situation | Action |
|---|---|
| Docs match knowledge | Proceed |
| Docs newer than knowledge | Proceed, cite fetched docs only |
| Can't fetch docs | HALT — tell human |
| Breaking changes found | HALT — show changes, ask human |
cat ~/.openclaw/openclaw.json # or whatever config file--- a/openclaw.json
+++ b/openclaw.json
@@ context @@
- "dmPolicy": "pairing",
+ "dmPolicy": "allowlist",
+ "allowFrom": ["tg:267378578"],For protected files, tell the sovereign:
# Apply on host via git
git apply /tmp/config-patch.diff
git diff
git commit -m "config: <description>"Never use `edit`/`write`/`apply_patch` on Tier 1 files.
Check a proposed config or diff against docs:
${VAR})"none" in productionOutput: VALID or INVALID with specifics.
Every change that gets applied must be logged:
### Change: [description]
- **Date**: [ISO 8601]
- **System**: [software + version]
- **Docs referenced**: [URL/path]
- **Proposed diff**: [summary]
- **Risk**: [Low/Med/High/Critical]
- **Approved by**: [name]
- **Rollback**: [steps]Store in: daily memory (memory/YYYY-MM-DD.md) + git commit message.
SOUL.md, USER.md, AGENTS.md, IDENTITY.md, core/, .env
openclaw.json, opencode.json, CLAUDE.md, GEMINI.md, docker-compose.yml
memory/.md, logs/.jsonl, skills/*/SKILL.md
| Risk | Examples |
|---|---|
| Low | Change model, add cron job |
| Medium | Change DM policy, add agent |
| High | Change auth, expose port |
| Critical | Rotate keys, delete data → 888_HOLD |
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.