blueprint-derive-rules
Pass
Audited by Gen Agent Trust Hub on Mar 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
Bashtool to executegitandjqcommands for inspecting repository history and updating project metadata. - [PROMPT_INJECTION]: The skill processes untrusted commit messages and descriptions, creating a surface for indirect prompt injection where malicious instructions could influence rule generation.
- Ingestion points: Commit logs are ingested via
git logcommands inSKILL.md(Steps 3, 4, and 5) andREFERENCE.md. - Boundary markers: There are no explicit delimiters or protective instructions used to separate the extracted commit data from the agent's internal logic.
- Capability inventory: The agent has access to powerful tools including
Bash,Write, andTask, which could be exploited if the agent obeys instructions embedded in commit logs. - Sanitization: The skill lacks evidence of sanitization or content validation for the commit messages before they are processed into rule statements.
Audit Metadata