spec-product-prototype
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface by processing data from external project files to generate documentation and trigger downstream routing logic.
- Ingestion points: The skill reads content from
{FEATURE_DIR}/requirements/prd.md,{FEATURE_DIR}/requirements/solution.md, and{FEATURE_DIR}/requirements/raw.md(specified in the '读取' section of SKILL.md). - Boundary markers: The instructions lack explicit delimiter requirements (e.g., XML tags or specific block markers) to isolate ingested file content from the agent's core instructions.
- Capability inventory: The skill has file-writing capabilities (creating
{FEATURE_DIR}/requirements/prototype.md) and invokes external routing logic via theusing-aidlcsub-skill, passing potentially influenced data in theROUTER_SUMMARYYAML. - Sanitization: No explicit sanitization or filtering is defined for the content extracted from the PRD, though the use of a fixed template (
assets/prototype-template.md) and the requirement for ASCII-only wireframes provide some structural constraints on the output.
Audit Metadata