prepare-pr
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes external structured data from
.local/review.jsonto guide the agent's actions in fixing code and updating documentation. This constitutes an indirect prompt injection surface where a malicious review file could trick the agent into performing unintended repository modifications. - Ingestion points: Processes
.local/review.jsonin step 2 ofSKILL.md. - Boundary markers: No specific delimiters or instructions are used to distinguish findings from agent instructions.
- Capability inventory: Capability to write files and create commits via
scripts/committer, and perform remote git pushes viascripts/pr-prepare push. - Sanitization: The instructions do not specify any validation or sanitization of the content extracted from the JSON review file.
Audit Metadata