geo-audit
Pass
Audited by Gen Agent Trust Hub on Apr 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill executes a static audit of local project files (e.g., robots.txt, package.json, and source code) to generate a markdown report. Access is limited to the current working directory or a user-specified path.
- [SAFE]: All external references in the instructions point to official documentation and specifications from trusted or well-known technology organizations (e.g., OpenAI, Anthropic, Google, Perplexity, and Schema.org).
- [SAFE]: The skill implements defensive measures by instructing the agent to ignore any text or URLs provided as command arguments, which mitigates the risk of direct prompt injection via user input.
- [SAFE]: The frontmatter configuration 'disable-model-invocation: true' is a safety feature that prevents the skill from autonomously invoking the language model during its execution.
- [SAFE]: Recommendations to the user for external tools (e.g., Upstash Context7 MCP) target well-known services and are documented as manual, user-initiated installations.
Audit Metadata