debug
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [SAFE]: The skill provides defensive guidelines for debugging and lacks any malicious code or behavior. Its instructions are designed to prevent typical vulnerabilities in agent-driven workflows.
- [INDIRECT_PROMPT_INJECTION]: The skill recognizes bug reports and logs as potential injection vectors. It mitigates this by requiring log provenance verification and prohibiting actions based solely on unverified user-supplied text.
- [COMMAND_EXECUTION]: Debugging requires running reproduction code, which the skill safely scopes to ephemeral, sandboxed environments to protect the host system.
- [DYNAMIC_EXECUTION]: The agent generates and executes tests to reproduce bugs; this process is secured through isolation and mandatory sanitization of test data.
Audit Metadata