motherduck-create-dive
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it ingests data from existing workspace Dives and guides that could contain malicious instructions designed to influence the agent's code generation.\n
- Ingestion points: Data is retrieved from the workspace via
read_dive,get_dive_guide, andlist_dives.\n - Boundary markers: No specific delimiters or instructions to disregard embedded commands are provided for data processed from workspace tools.\n
- Capability inventory: The skill can perform sensitive operations including writing to the workspace with
save_diveandupdate_dive, and executing SQL queries viamotherduck-query.\n - Sanitization: The instructions do not require the agent to sanitize or validate content fetched from the workspace before using it to generate or modify React and SQL components.
Audit Metadata