claude-code

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources including official documentation and GitHub issues. Analysis confirms the implementation of a sanitization pipeline (documented in the repository's reports) that defangs untrusted content by stripping HTML comments, imperative tags (e.g., , ), and wrapping external text in nonce-tagged blocks to prevent injection attempts.
  • [DATA_EXPOSURE]: The skill documents sensitive environment variables used by the Claude Code CLI (e.g., ANTHROPIC_API_KEY, CLAUDE_CODE_OAUTH_TOKEN) and provides instructions for safe credential management. No hardcoded secrets or unauthorized data access patterns were detected within the skill files.
  • [COMMAND_EXECUTION]: The documentation references various CLI commands and '!' shell-prefix syntax. It explicitly includes security warnings against shell injection, specifically advising against the interpolation of unsanitized arguments into shell commands, demonstrating a security-first approach to its documentation.
  • [EXTERNAL_DOWNLOADS]: The skill includes instructions for installing the Claude Code CLI and plugins from official sources (Anthropic's npm registry and GitHub). These references target well-known and trusted services, adhering to standard development practices for the documented tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 08:57 AM
Security Audit — agent-trust-hub — claude-code