skills/trailofbits/coop/integration/Gen Agent Trust Hub

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.sh to 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
Risk Level
SAFE
Analyzed
Sep 13, 2026, 09:49 PM
Security Audit — agent-trust-hub — integration