Most Movement Pack tasks supply destinations and status while the selected pathfinder owns routes, acceleration, obstacle avoidance, and physical movement. Perception tasks query the environment and return a condition instead of moving the agent.

The built-in navigation tasks use Unity’s navigation mesh. Integrations add variants for other supported pathfinding assets.

Configure a movement task

  1. Choose a task whose target relationship and completion behavior match the graph.
  2. Set a Transform or GameObject when the target can move. Use a position or rotation for a fixed value.
  3. Configure arrival distance, prediction, perception, or waypoint fields required by the task.
  4. Confirm the agent has the movement component required by that provider variant.

In Play Mode, watch task status. Navigation actions normally remain Running while moving and return Success after reaching their completion condition. Perception conditionals return Success only while their detection test passes.

Keep one movement authority

Do not let independent tasks, root motion, a Rigidbody controller, and navigation all write the same agent position or destination at once. The tree should select intent while one compatible movement implementation owns motion.

Choose a first task

  • Start with Seek when obstacles require pathfinding.
  • Use Move Towards only for unobstructed direct motion.
  • Use Patrol for authored routes.
  • Use Can See Object with a shared target variable before building a chase branch.

Verify the first behavior

Test success, failure, interruption, and an unreachable target. Compare the task arrival threshold with the movement component’s stopping distance before expanding the tree.

Continue with Task Reference after one Seek or Patrol action works reliably.