impeccable
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: A comprehensive review of the skill's 32 files revealed no malicious patterns, hardcoded credentials, or unauthorized network operations.
- [PROMPT_INJECTION]: The skill utilizes an indirect prompt injection surface in
references/teach-impeccable.mdthat instructs the agent to read project files (README, package.json) to establish design context. Evidence Chain: (1) Ingestion points: Project root files including README, package.json, and style guides; (2) Boundary markers: Absent; (3) Capability inventory: Read, Write, and Bash tools used for UI refinement; (4) Sanitization: Absent. This process is a core feature for establishing project-specific design principles and is not used for bypassing safety guidelines.
Audit Metadata