Skip to content

thread-tools.md

Latest commit

 

History

History
64 lines (45 loc) · 2.96 KB

thread-tools.md

File metadata and controls

64 lines (45 loc) · 2.96 KB
Raw

Agent thread tools

Agents can manage T3 Code threads in their current environment. A thread is a normal T3 Code conversation. It appears in the sidebar and stays available after the agent turn ends.

Available tools

  • create_thread creates a thread and sends its first message.
  • list_threads lists active threads in the current environment.
  • read_thread reads a thread and its recent history.
  • send_message_to_thread sends a message to another active thread.
  • wait_threads waits for threads to finish or need attention.

read_thread returns activity output only when the agent asks for it. It can use a cursor to read older history.

Threads and subagents

These tools create durable threads that you can open, manage, and continue. Each thread has its own conversation, project, model selection, permission mode, and workspace.

Native subagents are internal helper agents. T3 Code shows their task and transcript, but they do not create a new editable thread. Agents use native subagents for short, private fan-out work.

Scope and permissions

The tools can access only the calling thread's environment. They cannot read or control threads in another saved environment. A thread cannot send a message to itself.

Creating a thread or sending it a message can start agent work. A new thread inherits the calling thread's permission mode. A message sent to an existing thread uses that thread's current mode. Review Permission modes before you let agents create work in a shared checkout.

Thread tools use the same product MCP access as agent browser tools. In Settings, Integrations, turn on Agent browser access before you start a new agent session. A running agent keeps the tools it received when its session started.

Worktrees

An agent can create a thread in the current checkout or in a new Git worktree. A worktree uses a new branch and does not change the project's main checkout. T3 Code runs the project's configured new-worktree setup script there.

The agent can select a base branch. It can also ask T3 Code to start from the latest origin revision when the repository has that remote.

Limits

  • A prompt or message can contain up to 120,000 characters.
  • list_threads returns 50 threads by default and at most 200.
  • read_thread reads 10 turns by default and at most 50 at a time.
  • read_thread hides activity output by default. maxOutputCharsPerItem defaults to 8,000 characters per item and is limited to 20,000. It truncates returned message text. Activity payloads use the same limit but appear only when includeOutputs is true.
  • wait_threads accepts at most eight threads and waits up to five minutes. Its returned latestAssistantMessage is capped at 8,000 characters.

Set timeoutMs to 0 when an agent only needs the current thread state.

Provider availability

Thread tools are available in Codex, Claude Code, Cursor, Grok, and OpenCode sessions when T3 Code gives the session product MCP access.