#citizen

The stock citizen animgraph ships combat params that CitizenAnimationHelper does NOT wrap -- drive them with renderer.Set() directly, and replicate them yourself because the engine only networks locomotion.

The Citizen animgraph drives body aim from SetLookDirection("aim_body", dir, weight), not from the aim_body_pitch / aim_body_yaw floats. Writing only the floats reads back perfectly and moves the pose zero degrees.

HoldType shapes arm+fingers, body aim tilts via SetLookDirection (the aim_body_pitch/yaw floats are readback-only), and an IK hand target OWNS hand position, overriding the holdtype arm pose. Move the IK target to place a held item, use holdtype only for the fist.

The citizen's locomotion clips each carry their own authored ground speed in the animgraph's m_blendValue coordinates. Pacing every clip's PlaybackRate against one guessed nominal speed foot-slides every variant except the one it was tuned for.

Direct citizen sequence playback needs no AnimGraph, but SkinnedModelRenderer.UseAnimGraph must be set false BEFORE the first Sequence.Name write, or the write silently no-ops while the model's own animgraph keeps driving the pose. The citizen model exposes 466 sequences (raw directional locomotion + IdlePose family), enough for a full movement set with no rig work.

The Citizen ragdoll physics rig has 18 shapes (17 capsules plus 1 cylinder) and 16 joints (12 conical ballsockets plus 4 revolute hinges). A common folk estimate lands near 9 to 13 bodies, roughly half the real count, so it undercounts per-body and joint-solve cost. Both prefab files are plain KV3 text, so you count them offline with no editor running.