debugging
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze potentially untrusted data from external sources during the debugging process.
- Ingestion points: The agent is directed to read and analyze application logs, system logs (via
journalctl), and version control metadata (viagit diff,git log, andgit bisect). - Boundary markers: The instructions lack specific guidance on using delimiters or strict formatting to isolate external log/git data from the agent's primary instruction set.
- Capability inventory: The skill's workflow involves shell command execution (e.g.,
git,grep,journalctl), the ability to modify source code for isolation, and the execution of reproduction scripts and regression tests. - Sanitization: There are no instructions for sanitizing or validating the content of logs or git outputs before they are processed by the agent.
Audit Metadata