claude-code-agent-harness

Pass

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references a public GitHub repository (github.com/lintsinghua/claude-code-book) as a source for reference architecture and examples. This is a standard reference to external documentation hosted on a well-known service.
  • [DATA_EXFILTRATION]: Technical snippets in the documentation demonstrate architectural patterns for accessing the local file system and environment variables. These are provided as implementation examples for building agent capabilities such as memory systems and configuration loaders.
  • [PROMPT_INJECTION]: The documentation includes a conceptual design for a skill loader that uses template substitution to interpolate runtime variables into content. This is described as a system feature for creating dynamic agent prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 17, 2026, 06:51 PM
Security Audit — agent-trust-hub — claude-code-agent-harness