blast-radius
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requires the agent to verify code safety by dynamically generating and executing test scripts or code snippets in the local environment.
- Evidence: Step 5 of the 'Steps' section instructs the agent to 'Write a script or test that runs the real code, run it, and paste what happened.'
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted external content from pull requests and third-party libraries while maintaining the capability to execute code.
- Ingestion Points: The skill ingests untrusted data from diffs, PR commits, and library source code (Steps 1 and 3).
- Boundary Markers: There are no explicit instructions or delimiters defined to prevent the agent from following instructions embedded within the analyzed code diffs.
- Capability Inventory: The agent is granted the capability to write files, execute scripts, and perform symbol searches (Steps 3 and 5).
- Sanitization: No specific validation or sanitization of input code is required, although the skill includes a manual reminder to strip private information from final reports.
Audit Metadata