blast-radius
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to process untrusted external data, including code diffs, pull requests, and library source code, which could contain malicious patterns or instructions designed to influence the agent.
- Ingestion points: Pull requests and commits (via referenced tools), library source code, API JSON responses, and wire formats (Step 1, Step 3).
- Capability inventory: The skill directs the agent to write and execute scripts or tests that import and call the analyzed code (Step 4, Step 5).
- Boundary markers: The skill does not define specific delimiters or instructions to prevent the agent from following instructions embedded within the analyzed code.
- Sanitization: There are no explicit instructions to sanitize or validate the content of the external code before it is executed in a test script.
- [DYNAMIC_EXECUTION]: The skill explicitly commands the agent to generate and execute code at runtime to verify the safety of code changes.
- Evidence: "Write a script or test that runs the real code, run it, and paste what happened" (Step 5).
- Evidence: "Step 4 is usually one small script that imports the same library the app ships and calls the exact function you're worried about."
- [COMMAND_EXECUTION]: The instructions rely on the agent executing shell-level commands for searching and running verification code.
- Evidence: "The agent can grep those in a second."
- Evidence: "Look where grep stops."
Audit Metadata