rationale
Pass
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill mentions external services like Stripe, Auth0, Twilio, and SendGrid, and suggests searching public registries (npm, pip, crates, gems) as alternatives to building from scratch. These are used as illustrative examples of mature services and do not involve unauthorized downloads or executions.
- [COMMAND_EXECUTION]: The instructions reference internal agent tools such as
godmode:reference-engineandgodmode:github-search. These are part of the platform's expected capability set for locating code and documentation. - [PROMPT_INJECTION]: While the skill contains strong instructional language (e.g., 'No exceptions. No workarounds. No shortcuts.'), these are intended to define the agent's persona as a rigorous senior engineer. There are no attempts to bypass safety filters or ignore system instructions.
- [DATA_EXFILTRATION]: The skill processes user requests to generate an analysis. It does not access sensitive local files or transmit data to external third-party domains outside of its defined analytical purpose.
Audit Metadata