diagnosing-bugs
Pass
Audited by Gen Agent Trust Hub on Jul 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines a structured methodology for diagnosing bugs, emphasizing the creation of deterministic feedback loops and minimal reproductions. This is a standard and safe practice for software engineering agents.\n- [COMMAND_EXECUTION]: The skill instructs the agent to execute various debugging tools, including tests, curl commands, and headless browser scripts. These operations are intended for local development environments and are common for the described purpose.\n- [SAFE]: The instructions involve capturing and analyzing network traces and event logs. While these artifacts could contain sensitive information, the skill does not exfiltrate this data or access credentials.\n- [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes externally-provided error messages and log artifacts. However, the requirement for a deterministic feedback loop and the structured analysis phases mitigate the impact of potentially malicious data in these sources.\n
- Ingestion points: User-provided bug descriptions, captured trace files, and log dumps mentioned in SKILL.md.\n
- Boundary markers: None explicitly mentioned for external data.\n
- Capability inventory: Execution of CLI commands, tests, and browser automation tools.\n
- Sanitization: No explicit sanitization of error messages or trace content.
Audit Metadata