developing-genkit-dart
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides official installation instructions for the Genkit CLI via a shell script from the vendor-owned domain cli.genkit.dev and the Node Package Manager (npm). These resources are provided for official development workflows.
- [INDIRECT_PROMPT_INJECTION]: The skill guides the development of agents that ingest external data via the Model Context Protocol (MCP) and Agent-to-UI (A2UI) surfaces. 1. Ingestion points: Data enters the agent context through remoteAgent client calls, A2UI interactive surfaces, and MCP tool results. 2. Boundary markers: A2UI uses specific JSON envelopes and fenced code blocks to delimit instructions from prose. 3. Capability inventory: Agents are documented to use filesystem access, shell execution, and specialized tool middleware. 4. Sanitization: The documentation explicitly advises that renderers must sanitize inputs and that hosts should implement Content Security Policies (CSP) to mitigate risks.
- [DYNAMIC_EXECUTION]: The skill implements dynamic prompt generation using Dotprompt Handlebars templates and provides a custom agent API for orchestrating complex, multi-step workflows based on runtime model outputs.
Audit Metadata