write
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill's architecture creates a surface for indirect prompt injection by aggregating data from untrusted external sources and passing it to a writing agent.
- Ingestion points: Data enters the system through web search results in Step 2 (Agent 3) and React repository analysis in Step 2 (Agent 1).
- Boundary markers: Absent. The prompt template in Step 4 interpolates external research findings directly into the writing instructions without using delimiters (like XML tags) or providing 'ignore embedded instructions' warnings to the subagent.
- Capability inventory: The skill has the ability to create and modify documentation files in 'src/content/' and update navigation configuration files like 'sidebarLearn.json'.
- Sanitization: No sanitization, escaping, or validation logic is applied to the researched content before it is used to generate documentation.
Audit Metadata