# Openclaw Mem Self Model > Inspect and govern the experimental derived self-model sidecar. Use for continuity snapshots, attachment maps, drift, threats, public-safe wording, release receipts, or migration comparisons. - Skill: `phenomenoner/openclaw-mem-self-model` (Agent Skill) - Install (CLI): `npx skillmds@latest add phenomenoner/openclaw-mem-self-model` - Raw SKILL.md: https://api.skillmd.com/api/skills/phenomenoner/openclaw-mem-self-model/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: AI & ML - Author: phenomenoner (https://skillmd.com/u/phenomenoner) - Updated: 2026-09-10 - Page: https://skillmd.com/skills/phenomenoner/openclaw-mem-self-model --- # Self-model Lab The sidecar is derived, editable, and rebuildable. It is not consciousness, identity truth, or a second memory owner. Never write its claims back into Store truth automatically. ## Workflow 1. Build `current` for a scoped audit point. 2. Inspect attachment and adjudication before release decisions. 3. Check threat feed, explanation, and sensitivity before stronger wording. 4. Run wording lint before continuity-facing copy. 5. Use release only with a concrete operator reason; rebuild current to verify effect. ```bash openclaw-mem continuity current --scope --session-id --json openclaw-mem continuity attachment-map --snapshot --json openclaw-mem continuity adjudication --snapshot --json openclaw-mem continuity threat-feed --snapshot --json openclaw-mem continuity public-summary --snapshot --json openclaw-mem continuity wording-lint --snapshot --json openclaw-mem continuity release-history --scope --session-id --json ``` Persona priors are hints, not sovereignty. If evidence is thin or prior-dominated, report instability instead of inventing certainty. Enable autonomous receipts only when explicitly requested; use the governed soak controller under `` for long gates. ## Verify ```bash openclaw-mem continuity status --json python -m pytest tests/test_self_model_sidecar.py -q git diff --check ```