Back to skill
Skillv1.0.0
ClawScan security
MBTI · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
SuspiciousFeb 11, 2026, 9:03 AM
- Verdict
- suspicious
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- The skill claims to run a multi-stage MBTI diagnosis and automatically patch agent configuration, but the package only contains a no-op shell script and an instruction file that references many missing data files and an automatic SOUL.md modification — the implementation is incomplete and inconsistent with its claims.
- Guidance
- This skill is inconsistent: it promises a full MBTI diagnosis pipeline and automatic patches to SOUL.md, but the package only contains a stub script that echoes stage names and no data files or code that performs analysis or edits configuration. Do not rely on it to actually diagnose or modify an agent. Before installing or invoking it: (1) ask the author for the missing JSON/data files and the real implementation; (2) request a clear description of exactly what files/paths it will modify and preview the patches it would apply; (3) review the real code (or run it in a sandbox) to verify it doesn't exfiltrate data or change configs unexpectedly; and (4) only grant it the minimum privileges needed once you can inspect a complete implementation.
Review Dimensions
- Purpose & Capability
- concernThe SKILL.md describes a multi-stage diagnostic system and lists many core JSON files (questionnaires, personality mappings, diagnosis engine, config generator) and an automatic 'SOUL.md patches' capability. None of those data files or engines are present in the manifest. The only included code is a small shell script that only echoes stage names — it does not perform any diagnostics or apply configuration changes. This is a clear mismatch between stated purpose and actual capability.
- Instruction Scope
- concernInstructions tell the agent to run ~/.openclaw/skills/agent-mbti/scripts/run-diagnosis.sh full. The SKILL.md also describes automatic modification of SOUL.md (agent configuration). The provided script does not read or write any of the named JSON files or SOUL.md. The instructions therefore overreach: they promise reading/writing config and datasets that are not present and grant implied permission to change agent config even though no implementation is provided.
- Install Mechanism
- okNo install spec is provided and the skill is instruction-only with a tiny script — nothing is downloaded or installed. This limits immediate risk from supply-chain installs.
- Credentials
- noteThe skill declares no required environment variables or credentials (acceptable for a local diagnostic tool). However, the claimed ability to auto-modify SOUL.md (agent configuration) would require access to agent config paths/permissions; that capability is asserted but not implemented. If a future implementation added config-modifying code, required privileges should be clearly declared and justified.
- Persistence & Privilege
- okalways is false and there are no install hooks or self-modifying behavior in the current files. The skill could be invoked autonomously by the agent (platform default), which is normal; combined with the described but missing ability to edit SOUL.md this would be riskier, but the present package does not perform such modifications.