product-clarify
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill functionality is focused on local document management. It uses project-local scripts to automate the organization and metadata parsing of PDR files within a designated .adlc directory. No unauthorized command execution or data exfiltration patterns were identified.
- [INDIRECT_PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection as it is designed to read and interpret user-provided markdown documents (PDR-*.md). This is a fundamental aspect of its documented purpose for product decision management. 1. Ingestion points: Processes markdown files located in .adlc/drafts/pdr/ (referenced in SKILL.md and pdr-lib scripts). 2. Boundary markers: The skill does not implement explicit boundary delimiters or 'ignore instructions' directives when processing these files. 3. Capability inventory: The skill possesses the ability to read, write, and move files within the project repository using included library scripts (pdr-lib.sh and pdr-lib.ps1). 4. Sanitization: Metadata extraction is performed using standard text processing tools like awk and sed to isolate specific document sections.
Audit Metadata