review-for-openclaw
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external code and prompt changes that are untrusted, introducing a potential attack surface where malicious instructions could be embedded in the reviewed data.
- Ingestion points: Code or prompt changes provided in the OpenClaw workspace for audit or bug detection (SKILL.md).
- Boundary markers: The skill does not define explicit delimiters or instructions to treat the ingested code as data-only, increasing the risk of the agent obeying instructions hidden within the code.
- Capability inventory: While no specific scripts are provided, the skill is intended for use in an environment where the agent has workspace read and review capabilities.
- Sanitization: There are no sanitization steps, validation rules, or escaping mechanisms defined for the content being processed.
Audit Metadata