claude-code-ci

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the Claude Code CLI using npm install -g @anthropic-ai/claude-code. This targets the official package from Anthropic, which is a well-known and recognized service provider.
  • [PROMPT_INJECTION]: The skill illustrates how to ingest untrusted data (git diffs from pull requests) for automated review. This represents an indirect prompt injection surface where a contributor could embed instructions in a diff to influence the agent. The skill addresses this risk by demonstrating how to use --allowedTools to restrict the agent's capabilities to read-only operations during review tasks.
  • [CREDENTIALS_UNSAFE]: The integration examples correctly use GitHub Secrets (${{ secrets.ANTHROPIC_API_KEY }}) to handle authentication tokens, adhering to secure CI/CD configuration standards.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 12:52 PM
Security Audit — agent-trust-hub — claude-code-ci