claud
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: No malicious behavior or security vulnerabilities detected. The skill includes grounding rules to ensure responses are based on provided reference files.
- [EXTERNAL_DOWNLOADS]: The skill references official support and documentation URLs from Anthropic (e.g., anthropic.com, claude.com). These are trusted, well-known services.
- [SAFE]: The skill processes user queries about capabilities but does not utilize tools or perform actions that present a significant indirect prompt injection risk. Evidence: 1. Ingestion points: User queries in SKILL.md. 2. Boundary markers: Explicit grounding rules to use references/capability-matrix.md. 3. Capability inventory: No shell or network tools used. 4. Sanitization: Relies on model training.
Audit Metadata