afk-code-review
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from various sources which may contain malicious instructions intended to subvert the sub-agents' review process.\n
- Ingestion points: Untrusted data enters the agent context through
git diffoutput,git logcommit messages, PR metadata, and user-provided or repo-resident specification files (e.g., in.scratch/).\n - Boundary markers: The instructions for spawning sub-agents do not define strict delimiters (such as triple backticks or specific XML tags) to isolate the external content from the sub-agent instructions, increasing the risk of the agent obeying instructions embedded in the code or logs.\n
- Capability inventory: The skill uses sub-agents to perform analysis on the ingested data and aggregate findings; while this is a reporting role, manipulated output can mislead users.\n
- Sanitization: The skill does not implement any sanitization, filtering, or validation of the content retrieved from the git repository or local files before passing it to the sub-agents.
Audit Metadata