spec-product-clarify
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill implements strict logic gates, 'Hard Rules', and 'Red Flag' checks to ensure the agent does not skip clarification steps or produce routing conclusions prematurely, even when the user attempts to expedite the process.\n- [COMMAND_EXECUTION]: The skill invokes internal platform tools, specifically
spec-contextfor environment initialization andusing-aidlcfor automated workflow transitions based on a generated summary. These calls are integrated into the skill's operational logic.\n- [PROMPT_INJECTION]: The skill reads requirement data from untrusted sources, such as{FEATURE_DIR}/requirements/raw.md, and interpolates this content into structuredROUTER_SUMMARYYAML andsolution.mddocumentation. The lack of sanitization or boundary markers for this data creates an indirect prompt injection surface where malicious content in the requirement files could influence subsequent routing decisions or agent behavior.\n - Ingestion points:
{FEATURE_DIR}/requirements/raw.md,.aidlc/project/memory/product.md(SKILL.md)\n - Boundary markers: Absent; user-controlled data is integrated directly into prompts and structured tool inputs.\n
- Capability inventory: File system write access, execution of the
using-aidlcrouting tool (SKILL.md)\n - Sanitization: None detected for external requirement content.
Audit Metadata