LangSmith Fleet introduces two types of agent authorization: Assistants, which use the end user's own credentials, and Claws, which use a fixed set of credentials.

Langchain is a publication focusing on programming languages, language design, and compiler development. Readers can explore articles covering topics such as language features, syntax design, and compiler optimization techniques. Additionally, they can learn about programming language theory, language implementation challenges, and practical applications of language design principles.

LangChain

LangSmith Fleet introduces two distinct agent authorization models: Assistants, which act on behalf of end users using their own credentials, and Claws, which operate with a fixed set of credentials regardless of who interacts with them. The on-behalf-of model ensures users only access data they're permitted to see, while the fixed-credential model (inspired by patterns like OpenClaw) lets agent creators control exactly what resources the agent can reach. Both types support different communication channels (Slack, Gmail, Outlook, Teams), and fixed-credential agents are recommended to use human-in-the-loop guardrails for sensitive actions. Future work includes more granular memory permissions to prevent cross-user data leakage.

Two different types of agent authorization