document-editing
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill interacts with a document database via
pnpm actionscripts. These scripts perform structured operations like listing, searching, and updating documents using Drizzle ORM, ensuring raw SQL is not directly executed. - [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it reads and processes markdown content stored in a database.
- Ingestion points: Data is pulled into the agent's context through
get-documentandsearch-documentsinSKILL.md. - Boundary markers: The instructions do not specify any delimiters or safety markers to isolate retrieved document content from system instructions.
- Capability inventory: The agent has permissions to delete and update documents across the hierarchy, as well as add comments.
- Sanitization: No markdown sanitization or validation logic is mentioned in the skill instructions.
Audit Metadata