cursor-spec-interviewer
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements a robust 'Plan Mode' workflow that separates planning from execution, ensuring that repository changes are limited to approved documentation artifacts.\n- [SAFE]: Explicit safety rules are defined in the SKILL.md and rubric files to prevent the invention of API facts, the inclusion of secrets or credentials, and the execution of destructive migrations without explicit callouts.\n- [SAFE]: The workflow requires a mandatory user verification checkpoint to confirm scope, assumptions, risks, and validation plans before any files are saved to the repository.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from repository files and user requests. This surface risk is mitigated by explicit instructions to challenge assumptions against repo evidence and by requiring the user to verify all decisions before finalization.
Audit Metadata