claude-subagent-reference
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of documentation and reference material for the Claude Code subagent platform. All files were analyzed and found to be descriptive guides intended for information only.\n- [SAFE]: Analyzed for prompt injection; all instructions and examples (such as role-specific system prompts) are clearly presented as templates and do not attempt to override the host agent's safety protocols.\n- [SAFE]: Analyzed for data exposure and exfiltration; mentions of sensitive file paths (e.g., .env, .bashrc, or .aws/credentials) occur strictly within the context of explaining platform protection mechanisms and are not accessed or exfiltrated.\n- [SAFE]: The skill contains no active code, remote dependencies, or obfuscated content. Example scripts provided for feature demonstrations (like database query validation hooks) are security-focused and intended for local use by the user.
Audit Metadata