How to drive T3 Code threads from an outside agent
Copy a T3 Code MCP URL, sign Claude Code, Codex, or ChatGPT in with OAuth, and choose read-only or thread-driving access.
Copy a T3 Code MCP URL, sign Claude Code, Codex, or ChatGPT in with OAuth, and choose read-only or thread-driving access.
Pick subagents, agent view, teams, workflows, or Projects in Claude Code, and use the matching command to check work without mixing sessions.
How to define a Claude Code subagent in Markdown, choose project or user scope, and pass a session-only agent with --agents.
Enable Google Workspace MCP servers in Cloud, then connect Antigravity or Claude with OAuth so agents can use Gmail, Drive, and Calendar.
Point Claude, Claude Code, or Codex at Arkham's hosted MCP with an API key so agents can ask on-chain questions in plain English.
VS Code docs show how to add Agent Skills with a SKILL.md folder, where project and personal skills live, and how extensions contribute them.
LibreChat's docs cover declaring mcpServers in YAML and adding a server from the MCP Settings panel, including the OAuth callback.
Microsoft's APM guide scaffolds a package, adds a skill and custom agent, and installs them into Copilot with apm install.
The MCP docs walk through adding a Claude custom connector to a remote server URL and completing that server's authentication.
Add a pre-configured or custom MCP server to a Notion Custom Agent, with per-agent auth and confirm-on-write defaults.
Install Pydantic AI, create an Agent, and run it synchronously, asynchronously, or as a stream.
Grant IAM, enable a Model Garden API Service card, and call a managed open model on Gemini Enterprise Agent Platform with the Gen AI or OpenAI SDK.
Create a Claude Code skill as a SKILL.md folder, place it in personal or project skills, and invoke it with /skill-name or by description.
OpenAI's Oct 2 GPT-6 guide covers model choice, reasoning effort, caching, and how to keep long agent runs moving.
Cloudflare's Agents docs show the starter commands to run a durable chat agent locally on Workers AI, with tools and approvals.
Point OpenClaw at the hosted Agents API harness and run a code task from chat without provisioning a separate execution machine.
Connect an existing agent to Microsoft Agent 365 with skills that register identity, observability, and governed Microsoft 365 tools.
Create an OpenAI Agents API session that writes and runs a script in a hosted sandbox, then continue or delete that session.
Advertise an events capability and send webhook updates so ChatGPT can subscribe to real-time changes from an OpenAI plugin MCP server.
Official OpenAI Agents SDK docs show how to name traces, hide sensitive span data, flush worker exports, and add or replace processors.
Sep 28 DataCamp walkthrough scores agent traces on task completion, tool selection, argument validity, and step budget, not final-answer LLM scores.
Install the prerelease Blazor AI package, bind ChatPage to a UIAgent, and connect remote agents through the AG-UI .NET client.
How to use Claude Code's /claude-api build-eval and hillclimb commands, including a held-out test set the hillclimber never sees and a revert when train improves and test is flat.
Install Foundry Dev Pack, initialize a hosted agent with azd, run it locally, and deploy it to Foundry Agent Service using Microsoft’s current quickstart.
BMC’s Control-M SaaS docs show how to enable the preview MCP server and connect Claude Desktop, VS Code, or Cursor.