debugging
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing external, potentially untrusted data such as bug reports, log files, and request payloads to facilitate debugging.\n
- Ingestion points: User-provided failure reports, triggers, and inputs are processed in
SKILL.mdandreferences/debug-loop.md.\n - Boundary markers: The skill instructs the agent to define failures and build reproducers, which creates a scope for analysis but does not include explicit instructions to disregard embedded malicious prompts in the source data.\n
- Capability inventory: The skill uses
BashandWritetools to execute reproducers and verify fixes, providing a vector if ingested data contains command injection or malicious instructions.\n - Sanitization: No specific sanitization or filtering of external inputs is described.\n- [SAFE]: The skill uses well-known debugging tools like
git bisectand provides guidance for handoffs to other specialized skills.\n- [SAFE]: External documentation links point to trusted organizations and well-known services including Microsoft, Google, and GitHub.
Audit Metadata