TASK FORCE DEPLOYMENT GUIDE

When you deploy a Task Force, you are triggering a highly concentrated, multi-agent cognitive loop. Your mission parameters are simultaneously fed to three distinct AI models, which then execute a strict, three-round processing sequence.

You control the physics of the interaction by selecting the Mode:

  • Adversarial Mode: The agents are programmed for maximum friction. They will aggressively audit your premises, hunt for logical fallacies, and attempt to destroy weak architecture. Use this to red-team a strategy before deploying it in the real world.
  • Symposium Mode: The agents are programmed for collaborative synthesis. They will build upon each other’s logic, filling in gaps and constructing a cohesive, finalized blueprint.

The Physics of the Prompt (Variable Overload)

To extract the maximum value from this engine, you must understand its mechanical limits.

The Task Force operates on a strict three-round cycle. If you feed the Task Force a massive, unstructured mission with 15 million unresolved variables (e.g., a raw, “vibe-coding” start), the agents will be forced to spend all three rounds simply trying to define the parameters of the problem. A prompt that is too broad forces the AI into shallow summarization.

Best Practices: The Surgical Strike

The narrower the mission and the fewer the variables, the deeper the engine cuts. The Task Force yields the most devastatingly effective results when used to solve the exact point of failure in a nearly finished concept.

For maximum structural integrity, utilize the following workflow:

  1. 1
    The Rough Draft (1:1 Chat): Use the standard “Chat with Logic Trace” to brainstorm your raw ideas. Use the 1:1 agent to build your outline, write your initial code, or draft your business proposal.
  2. 2
    Isolate the Bottleneck: Once you have a working draft, identify the specific mechanism that you cannot make work (e.g., “The pricing model has a loophole,” or “This Python routing function keeps crashing”).
  3. 3
    Deploy the Task Force: Feed that specific, isolated bottleneck to the Task Force.

While the Task Force engine possesses the raw compute power to handle highly complex, unstructured issues from a cold start, treating it as a surgical instrument rather than a blunt club will consistently produce superior, mathematically unassailable results.

The Final Directive: Preliminary Red-Teaming

Never send raw, untested assumptions into a multi-agent crucible. Before deploying a Task Force, subject your initial premise to a preliminary stress-test in the 1:1 Chat. By actively red-teaming your own variables and pruning the obvious logical fallacies first, you deliver a purified payload to the Task Force.

This allows the networked agents to bypass the baseline noise and dedicate 100% of their compute exclusively to the hardest structural constraints of your problem.

For a precise mechanical breakdown of this protocol, see Red Teaming in the Docs.