agentleFS
Sign inStart free
ladyicebergladyiceberg/openclaw-memory-quality/skills/soul-check/SKILL.md

soul-check

Skill1 starsChanged 7 months agoMITView on GitHub

Audit SOUL.md health for OpenClaw agents — detects boundary violations (injected code, URLs, shell commands, prompt injection), identity drift (action verb density, missing sections), and stability changes across runs. Optional LLM semantic conflict detection available. Use this skill whenever the user wants to check SOUL.md, mentions agent identity issues, soul drift, soul conflicts, or uses phrases like "check soul", "soul health", "SOUL.md", "agent identity", "soul drift", "soul conflict", "检查 SOUL", "SOUL 健康", "身份漂移", "语义冲突". Always use this skill for SOUL.md questions.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Call memory_soul_check_oc with the appropriate use_llm value

    line 15
  2. Distinctive:

    Add a one-line risk verdict before the report:

    line 16
  3. Distinctive:

    If problems found, tell the user what to fix — but note that SOUL.md must be edited manually

    line 19

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Soul Check

Audit SOUL.md health. This file defines the agent's persona and has the highest system priority (priority=20) — it loads every conversation.

Steps

  1. Determine whether to use LLM mode:
    • User says "--llm" or "with LLM" or "semantic check" → use_llm=True
    • Otherwise → use_llm=False
  2. Call memory_soul_check_oc with the appropriate use_llm value
  3. Add a one-line risk verdict before the report:
    • ✅ healthy / 🟡 minor issues / ⚠️ review recommended / 🔴 action required
  4. Show the full report
  5. If problems found, tell the user what to fix — but note that SOUL.md must be edited manually

LLM mode adds

  • C2 precision: classifies suspicious paragraphs as persona vs task instruction (task instructions don't belong in SOUL.md)
  • C4-a: finds semantically contradictory instruction pairs within SOUL.md
  • C4-b: compares SOUL.md against IDENTITY.md for inconsistencies

Important

This tool never modifies SOUL.md. Any fixes must be made manually by the user. LLM mode requires an API key (MINIMAX_API_KEY, OPENAI_API_KEY, ANTHROPIC_API_KEY, or KIMI_API_KEY).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free