do-create-techspec

Pass

Audited by Gen Agent Trust Hub on May 8, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill performs a deep analysis of the project codebase in Step 3 and subsequently uses Web Search in Step 4. There is an inherent risk that project-specific details, architectural decisions, or code snippets discovered during analysis may be included in search engine queries, leading to data exposure.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes untrusted data from user-provided PRD files and the local codebase.
  • Ingestion points: Files located at prds/prd-[feature-slug]/prd.md and any file discovered during project exploration in Step 3.
  • Boundary markers: None identified. The instructions do not specify delimiters or warnings for the agent to ignore embedded instructions in the ingested files.
  • Capability inventory: The skill uses tools for directory listing, file reading, file writing to ./prds/prd-[feature-slug]/techspec.md, Web Search, and MCP tool invocation (Context7).
  • Sanitization: No sanitization or validation logic is defined for the content extracted from the PRD or codebase before it is used to generate the final specification.
Audit Metadata
Risk Level
SAFE
Analyzed
May 8, 2026, 02:50 AM