bridge-review
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill instructions (Step 3) direct the agent to 'write and run temp script' to verify review findings. This involves creating and executing code generated during the session.\n- [INDIRECT_PROMPT_INJECTION]: The skill relies on output from peer agents via the 'agent-bridge' tool to determine fixes and tests.\n
- Ingestion points: Data received from the 'agent-bridge' tool containing peer review findings.\n
- Boundary markers: The skill lacks explicit delimiters for processing the peer agent's output.\n
- Capability inventory: Includes file reading, file writing (Step 4), and command execution for tests and verification scripts (Step 3).\n
- Sanitization: Relies on the agent's internal verification step to check findings against the code before acting.\n- [COMMAND_EXECUTION]: The agent is instructed to 'run the test' and 'try the input', necessitating the execution of shell commands within the environment.
Audit Metadata