blast-radius
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data such as code diffs and library source code (SKILL.md, Step 1 and 3). This ingestion point lacks explicit boundary markers, potentially allowing malicious content within the analyzed code to influence the agent. The skill has capabilities for filesystem access and script execution (SKILL.md, Step 5). It includes a sanitization step to strip private information before sharing output.
- [COMMAND_EXECUTION]: The instructions direct the agent to run command-line searches (grep) and execute verification tests (SKILL.md, Step 4 and 5).
- [DYNAMIC_EXECUTION]: The agent is instructed to write and execute scripts that import local libraries to prove safety facts through runtime behavior (SKILL.md, Step 5).
Audit Metadata