debug-mode
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- NO_CODE (SAFE): All files (SKILL.md, agents/debug-specialist.md, and commands/debug.md) are Markdown documents containing instructional text. No Python, Node.js, or shell scripts are included.\n- PROMPT_INJECTION (SAFE): The skill's instructions do not contain patterns typical of prompt injection or jailbreak attempts. It provides standard operational procedures for debugging without attempting to bypass safety filters or override system-level instructions.\n- Indirect Prompt Injection (SAFE): The skill defines a workflow for analyzing external data (logs and reproduction steps), which is a common surface for indirect prompt injection. However, as the skill provides no code to implement this analysis, it does not facilitate an attack. 1. Ingestion points: User-provided error logs and stack traces in the Reproduction and Analysis phases. 2. Boundary markers: None specified. 3. Capability inventory: Code exploration and targeted fixes (file system access). 4. Sanitization: None mentioned. The risk remains at the base model's safety layer.
Audit Metadata