Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Structured identity graph system for AI agents. Reads agent memory and soul files, extracts beliefs, traits, relationships, and contradictions, then produces versioned snapshots with a numeric continuity score. Tracks identity drift over time and supports a freeze protocol before model upgrades.
Quantifies identity stability with a scored metric (KL divergence + MSE) rather than relying on qualitative inspection of raw memory files.
When to use it
- Tracking belief drift across agent versions
- Freezing agent identity before a model upgrade
- Comparing personality trait snapshots between runs
- Detecting continuity fractures in long-running agents
- Auditing how an agent's core values shift over time
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Run a full identity cycle on my agent
- 1Reads MEMORY.md and SOUL.md from the agent workspace
- 2Calls Gemini API to extract beliefs, traits, relationships, cognitive patterns, and contradictions
- 3Computes Continuity Score using KL divergence on beliefs and MSE on traits
- 4Saves structured identity graph to current_identity.json
- 5Writes versioned snapshot and diff to snapshots/ and diffs/
Versioned identity snapshot with a continuity score, e.g. 0.87 (drift range: 0.75-0.89)
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.