cursor-spec-interviewer

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's primary function is generating documentation and implementation plans based on user input and existing repository context. Analysis of the 14 files reveals no malicious patterns, obfuscation, or unauthorized access to sensitive data.
  • [COMMAND_EXECUTION]: The skill facilitates the execution of local repository validation tools (e.g., pnpm test, pnpm lint) as specified in the generated implementation spec. These executions are intended for verification and are based on existing project configurations.
  • [EXTERNAL_DOWNLOADS]: The instructions permit the agent to use web search or MCP tools to verify technical details in official external documentation. This is restricted to informational purposes and uses the agent's secure toolset.
  • [PROMPT_INJECTION]: The skill is exposed to indirect prompt injection (Category 8) by reading untrusted repository context (Ingestion: README.md, .cursor/rules, etc.). It employs boundary markers (Boundaries: <implementation_spec> tags) and utilizes capabilities for file writing and executing local validation commands. Sanitization is achieved through a mandatory human-in-the-loop verification checkpoint before any artifacts are persisted.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 01:05 PM
Security Audit — agent-trust-hub — cursor-spec-interviewer