blast-radius
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content including code diffs, PR comments, and library source code which may contain malicious instructions intended to manipulate the agent's behavior.
- Ingestion points: The agent reads diffs, symbols, and PR/commit data (Step 1), and analyzes external library source code (Step 3).
- Boundary markers: The instructions lack explicit delimiters or warnings to treat analyzed code as data rather than instructions, potentially leading to accidental obedience of embedded commands.
- Capability inventory: The agent is explicitly instructed to write and execute scripts or tests (Step 5) to prove safety facts.
- Sanitization: No sanitization or filtering of the analyzed code is specified before it is processed or used to inform script generation.
- [COMMAND_EXECUTION]: The skill workflow requires the agent to write and execute scripts or tests that call project code. While intended for legitimate testing, this involves the execution of agent-generated code within the local environment.
- [DYNAMIC_EXECUTION]: The skill involves dynamic code generation where the agent creates new scripts at runtime (Step 5) to validate assumptions about the codebase.
Audit Metadata