comfy-custom-nodes-search
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external repository content through a code search tool.
- Ingestion points: Untrusted results from
comfy --json code-searchare brought into the agent's context. - Boundary markers: None specified. The instructions lack technical delimiters (e.g., XML tags or markers) to isolate external content from the agent's instructions.
- Capability inventory: The search results are used to influence the agent's code generation, API implementation, and dependency management.
- Sanitization: Not present. While the skill warns that results are untrusted and stale, it does not provide mechanisms for the agent to sanitize or validate the external data before use.
Audit Metadata