skills/vercel/chat/chat-sdk/Gen Agent Trust Hub

chat-sdk

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [External Package References]: The skill mentions the installation of the chat npm package and the use of the create-chat-sdk CLI tool.
  • Using external packages and CLI tools is a standard part of modern development workflows. It is a best practice to verify package integrity and ensure you are using the intended versions from reputable registries.
  • Context: These tools are presented as the primary means to scaffold and build projects using the SDK described in the skill.
  • [Local Data Ingestion]: The instructions direct the agent to read documentation and resource files located in node_modules/chat/docs/ and node_modules/chat/resources/ to gain technical context.
  • Ingesting content from external files creates an indirect prompt injection surface. If these files were to contain instructions, they could potentially influence the agent's behavior during the session.
  • Context: This is a common pattern for AI-assisted development tools to provide agents with the specific domain knowledge required to assist with a particular library. Users should ensure their local environment contains legitimate versions of the referenced resources.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 02:04 PM