skill-architecture

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as a documentation-heavy guide for developers, focusing on architectural patterns and best practices for the Claude Code platform.
  • [SAFE]: Includes a dedicated security reference (references/security-practices.md) that educates users on defending against prompt injection, tool abuse, and data exfiltration.
  • [SAFE]: Recommends the use of 'allowed-tools' to restrict agent capabilities, emphasizing a least-privilege security model.
  • [SAFE]: Provides guidance on safe path handling and bash compatibility to prevent common execution errors and security pitfalls in diverse environments.
  • [SAFE]: All external URL references are to official documentation (Anthropic), well-known specification sites, or trusted repository sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 11:47 PM