evidence-first-debugging
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill provides a methodological framework for software debugging and incident response. It emphasizes objective evidence collection and requires falsifiable tests for all hypotheses, which promotes safer and more accurate agent behavior.
- [NO_CODE]: The skill consists exclusively of instructional Markdown content and Mermaid diagrams. It does not include executable scripts (e.g., Python, JavaScript, Shell) or configuration files that initiate external downloads or command execution.
- [INDIRECT_PROMPT_INJECTION]: The skill is intended to process external data such as software logs, code, and system outputs, which are potential vectors for indirect prompt injection. However, the skill implements structural defenses by requiring every factual claim to be cited with an Evidence ID and mandating explicit disclosure for truncated data, which helps maintain context boundaries.
Audit Metadata