genkit-integrations-js
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill relies on fetching external documentation via the
search_genkit_docsandread_genkit_docstools to guide its auditing and code generation workflows. This creates a vulnerability surface where instructions embedded in the processed data could influence agent behavior. - Ingestion points: External content is ingested through the
read_genkit_docstool call described in the Source of Truth section of SKILL.md. - Capability inventory: The agent is authorized to generate implementation code, choose documented primitives, and produce audit findings based on the fetched data.
- Boundary markers: The skill lacks instructions to treat external data as untrusted or to use specific delimiters to isolate embedded instructions.
- Sanitization: There are no verification or sanitization steps mentioned to validate the integrity of the documentation before the agent acts on it.
Audit Metadata