cursor-spec-interviewer
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill instructions are focused on structured planning and do not contain attempts to bypass safety filters or override agent behavior. It defines clear goals and constraints for the interview process.
- [DATA_EXPOSURE_AND_EXFILTRATION]: No sensitive file paths or hardcoded credentials were found. Network operations are described only for fetching documentation from official sources using standard tools.
- [OBFUSCATION]: The skill content is clear and human-readable. No Base64 encoded payloads, zero-width characters, or hidden text patterns were detected.
- [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill does not install external packages or execute remote code. It generates documentation artifacts and suggests standard local validation commands.
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect injection as it ingests content from repository files.
- Ingestion points: User inputs and repository configuration files (SKILL.md, README.md, .cursor/rules).
- Boundary markers: The output templates in assets/cursor-execution-prompt.md use specific XML-like tags to delimit generated content.
- Capability inventory: The skill performs file system writes to persist specs and ADRs (SKILL.md).
- Sanitization: The workflow includes a mandatory user verification checkpoint before final artifact creation.
- [DYNAMIC_EXECUTION]: No evidence of runtime code generation, unsafe deserialization, or process injection. The skill generates static Markdown files based on provided templates.
Audit Metadata