integration
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates the execution of
./tests/run-integration.shto run integration tests. It allows passing arguments for local execution on Lima or remote execution on Firecracker. This command execution is consistent with the skill's stated purpose of integration testing.\n- [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to redirect the test suite output to a file and then inspect the full content to report results. This establishes an ingestion point for untrusted data generated by the test process.\n - Ingestion points: Output file generated by the integration runner in SKILL.md.\n
- Boundary markers: None; the agent is instructed to inspect the 'complete output'.\n
- Capability inventory: The agent can execute shell scripts and perform file system operations (read/write/delete).\n
- Sanitization: No sanitization or filtering of the command output is requested before interpretation.
Audit Metadata