tui-prd-to-descriptions

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its core function of processing external data to generate instructions for other agents.
  • Ingestion points: Processes user-provided PRD documents, interface descriptions, and functional requirements (defined in SKILL.md).
  • Boundary markers: Absent. There are no instructions to use delimiters or ignore embedded commands within the input PRD text to prevent it from influencing the output prompts.
  • Capability inventory: While this skill does not directly invoke tools (confirmed by its documentation), it generates high-privilege plans (PENCIL_PLAN) containing tool calls such as mcp__pencil__batch_design for downstream consumption by pencil-skills or other agents.
  • Sanitization: No sanitization or escaping of the input content is performed before it is interpolated into the generated STITCH_PROMPT or PENCIL_PLAN outputs.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 04:07 PM
Security Audit — agent-trust-hub — tui-prd-to-descriptions