skills/davekilleen/dex/product-brief/Gen Agent Trust Hub

product-brief

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill implements a workflow that captures untrusted user input and incorporates it into files saved on the local filesystem, which presents a surface for indirect prompt injection.
  • Ingestion points: The skill captures user input in SKILL.md during the initial idea phase and throughout the conversational questioning process in Phase 2.
  • Boundary markers: While the skill uses Markdown structural templates, it lacks explicit delimiters or specific instructions to the agent to treat interpolated user content strictly as data rather than potential instructions.
  • Capability inventory: The skill performs multiple file-write operations, creating project folders, saving PRD documents in 04-Projects/, and updating stakeholder and task files.
  • Sanitization: There is no evidence of input validation or sanitization routines to filter potentially malicious instructions within the user's product descriptions before they are written to files.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 12:41 AM
Security Audit — agent-trust-hub — product-brief