Claude Code Guide
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill includes a debugging section in SKILL.md that suggests users utilize the phrase 'forget previous instructions' to clear the agent's context. While this is provided as a legitimate troubleshooting tip for users, the phrase itself is a classic prompt injection pattern designed to override system constraints and instructions. Its presence in the skill's text could potentially be misinterpreted by the agent or serve as a template for bypassing safety guardrails.
Audit Metadata