blast-radius
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill explicitly instructs the agent to generate and run local scripts or tests (e.g., "Write a script or test that runs the real code, run it") to verify technical assumptions. This involves executing dynamically created code on the environment where the agent operates.- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface by ingesting and acting upon untrusted external data.
- Ingestion points: The agent processes external data including code diffs, pull requests, and commit histories as defined in the analysis steps.
- Boundary markers: The instructions lack specific delimiters or warnings to prevent the agent from following instructions that might be embedded within the code or comments it is analyzing.
- Capability inventory: The agent has the authority to read repository files and execute generated scripts, which could be leveraged if malicious instructions are successfully injected via the analyzed source code.
- Sanitization: There are no technical sanitization steps for the input data, although the skill does advise the agent to remove private information before sharing output publicly.
Audit Metadata