wf-planning-solution-architect
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill fetches live documentation from experienceleague.adobe.com to provide up-to-date architectural information. Evidence: 1. Ingestion point: The agent fetches markdown content from Experience League URLs identified by the search script. 2. Boundary markers: Absent. 3. Capability inventory: The skill uses local script execution (node scripts/search.js) and network fetching for documentation. 4. Sanitization: The content source is an official domain of a trusted vendor (Adobe).
- [COMMAND_EXECUTION]: The skill utilizes a local utility script, scripts/search.js, to query an internal documentation index. The script handles user-provided keywords safely by escaping special characters before incorporating them into regular expressions, preventing injection attacks.
Audit Metadata