interrogate
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data which could be manipulated to influence the agent's behavior during the testing phase.
- Ingestion points: The skill derives its 'intent' from external sources such as 'the user's message, the commits, and the PR description' in SKILL.md.
- Boundary markers: There are no explicit markers or instructions provided to the subagents to ignore potentially malicious content within the analyzed scope.
- Capability inventory: The skill creates 'git worktree' environments and executes shell commands or repro scripts to verify potential code failures.
- Sanitization: The skill does not implement validation or escaping for the data ingested from commit messages or PR descriptions.
- [COMMAND_EXECUTION]: The skill performs shell command execution to create isolated environments and run tests or reproduction scripts.
- Evidence: SKILL.md instructs the use of 'git worktree' and the execution of commands like 'codex exec', 'claude -p', and 'repro scripts' to verify findings against the codebase.
Audit Metadata