spec-writer
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from multiple sources to generate product requirements.
- Ingestion points: The skill reads strategic project context from local files in
docs/strategic/(e.g.,BUSINESS_CASE.md) and accepts free-text feature descriptions from the user in Step 1. - Boundary markers: Instructions do not specify explicit delimiters or isolation markers for the ingested data during the generation phase, relying instead on the structured PRD template.
- Capability inventory: The skill possesses the ability to read existing project files, write new files (
PRD.md), and trigger external logic viasdlc-checkpoint.skill. - Sanitization: The skill explicitly implements a 'Requirement Sanitization' guardrail designed to scrub hardcoded credentials or malicious patterns from user inputs. It also requires a human-in-the-loop approval step before the PRD is finalized or implemented.
Audit Metadata