managed-deep-agents

Your first hosted agent, shaped, tested, and deployed.

AI agent development

What it does

Builds, tests, and deploys a code-first Managed Deep Agent after matching the user's requirements to supported capabilities.

When to use it

Use when building a code-first Deep Agent on LangSmith or adding managed capabilities to one.

Tell it what the agent should do and who or what will interact with it; it confirms the design, creates the project, tests it, and deploys it.

What you provide

This skill

Context Hub

Write

Syncs and deletes Context Hub files (irreversible)

LangSmith

Write

Uploads source and hosted secrets

LangSmith deployment

Execute

Deploys the agent to LangSmith

LangSmith deployment dashboard

Read

Reads deployment builds and traces

Requires a LangSmith workspace with Managed Deep Agents public beta access.

The workflow reads the selected model provider's API key from the project .env after the user supplies it.

Docker is optionally required for Harbor evals using its default environment.

Slack channel support optionally reads the signing secret from the project .env.