solicit-gemini-review
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local script
./scripts/agents/tooling/agentTool.tsto interact with pull request metadata and trigger reviews. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it ingests and processes content from pull request reviews and comments. Ingestion points: Pull request info and review threads are fetched via
agentTool.tsin SKILL.md. Boundary markers: No explicit delimiters or instructions to ignore embedded commands were found. Capability inventory: The skill can tag pull requests and generate reports usingagentTool.ts. Sanitization: There is no evidence of sanitization for the data retrieved from external review threads.
Audit Metadata