Long Horizon & Model Council

Most tasks should stay as Standard tasks. They are simpler, faster, and easier to steer.

When a task needs more structure, Meowbert now offers two workflow modes in the Task Parameters menu:

• Long Horizon — for tasks that need a durable plan plus explicit review before they are allowed to finish.
• Model Council — for tasks that benefit from multiple agents discussing the work together before a final answer is delivered.

You choose the workflow before sending the task.

Long Horizon

Use Long Horizon when you want the agent to stay disciplined over a larger piece of work.

What happens

1. A clarify stage runs first instead of starting execution immediately.
2. The agent asks clarifying questions if needed.
3. Once ready, it writes a durable PLAN.md.
4. The main agent executes against that plan.
5. When it thinks it is done, it submits its work to a reviewer.
6. The task only finishes once the reviewer approves.

Why use it

Long Horizon is a good fit when:

• the task is broad and easy to half-finish,
• quality matters more than speed,
• you want a visible plan before deep execution starts,
• you want review pressure before the final answer is sent.

What you see in Task Detail

The workflow panel shows:

• the saved plan,
• the latest submission,
• the current round,
• the reviewer verdicts for that round.

If a reviewer wants changes, the task keeps going instead of silently stopping early.

Model Council

Use Model Council when one agent should not be the only voice working on the task.

What happens

1. Meowbert creates one leader plus 2 to 8 workers.
2. The leader starts the shared discussion.
3. Workers collaborate through shared council channels.
4. The leader is responsible for the final user-facing answer, but only after the council has discussed the work.

This is intentionally different from subtasks. Subtasks are independent. A Model Council is a shared discussion.

Why use it

Model Council is useful when:

• the task is ambiguous and you want multiple viewpoints,
• you want parallel investigation plus debate,
• verification matters and it is healthy for agents to challenge each other,
• you want to inspect how different workers approached the same problem.

Configuring the council

In the composer, you can choose the worker count:

• minimum: 2
• maximum: 8

The leader is added automatically.

Tool access

Whatever tools you enable in the composer apply to the council too. If you enable Web Search, Memory, Computer Use, Subtasks, or Skills, the council agents receive those same selected capabilities for their runs.

What you see in Task Detail

The workflow panel shows:

• the worker list,
• the channel list,
• live channel popups,
• each worker's live POV using the normal task detail renderer.

This makes it much easier to inspect who said what and what each worker is currently doing.

Choosing the right mode

Use Standard task when:

• the task is straightforward,
• speed matters,
• one agent is enough.

Use Long Horizon when:

• you want a plan first,
• you want formal review before completion,
• you are worried about the agent stopping too early.

Use Model Council when:

• you want discussion between multiple agents,
• you want disagreement and verification,
• the task benefits from several concurrent perspectives.

Important limits

• Workflow type is chosen at task creation time.
• After the task is created, the workflow type is fixed.
• You can still adjust normal task parameters like max steps or waiting behavior where supported.