principles

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is primarily instructional, providing best practices for code editing and project management within a development environment.
  • [DATA_EXPOSURE]: The repository exploration module (references/repo-context-reading.md) includes instructions to read common project files like README.md and package.json. It explicitly contains a safety directive to avoid reading sensitive directories and files such as .env or secrets/.
  • [COMMAND_EXECUTION]: The skill suggests standard, non-malicious shell commands for version control status (git status, git diff) and build verification (npm run build, npx tsc). These are appropriate for the skill's stated purpose of guiding development workflows.
  • [PROMPT_INJECTION]: The vibecoder-guide.md provides response templates for the agent to use when interacting with users. These templates are informative and do not contain any instructions to bypass safety filters or override system constraints.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 05:15 PM
Security Audit — agent-trust-hub — principles