Animation toolsets

Animation is the one domain where the honest answer is that ClaudusBridge is not there yet. The agent can read an Animation Blueprint deeply enough to explain why a character is stuck in the wrong state, and it can repair one specific class of broken binding. It does not provide a supported AnimGraph-authoring or IK Rig/IK Retargeter workflow. Generic reflection does not establish those domain-specific editing guarantees. If your reason for buying is an animation pipeline, read this page before you decide.

What works today

ClaudusBridge ships a native AnimBlueprint toolset of five tools, written because the generic Blueprint graph tools read K2 event and function graphs but are blind to AnimGraph state-machine topology.

• DescribeAnimStateMachines dumps every state machine of an Animation Blueprint: states with the animation assets they play, the entry state, conduits and aliases, and every transition edge with its source, target, priority and automatic-rule flag, plus a summary of the rule graph including the Property Access paths, variable reads and function calls each rule evaluates.
• DescribeTransitionRule dumps the full rule graph of one transition edge, node by node, with classes, titles, pins, default values and connections.
• DescribeAnimGraph dumps every node of a named graph including the root AnimGraph itself, which the generic Blueprint graph tools cannot read, with each node's resolved Property Access path.
• RebindAnimPropertyAccess is the one write primitive: it rewrites the dotted path of every Property Access node across all graphs, transition rules included, whose current path matches a substring, then recompiles and optionally saves. It exists to repair stale bindings after a variable or component rename.
• GetActiveAnimStates reports the live state of a running actor, preferring a PIE world: per skeletal mesh component, the anim instance class and the currently active state of each state machine. That lets an agent verify animation behaviour headlessly instead of asking a human to watch the viewport.

What is not supported yet

• IK Retargeter. There are no tools to create, inspect or edit IK Rig or IK Retargeter assets, and no way to configure a retarget chain or bone mapping.
• Retargeted assets. The agent cannot run a batch retarget to produce retargeted animation sequences from a source skeleton to a target one.
• AnimBP authoring. It cannot create an Animation Blueprint, add or remove states, add transitions, author transition rule logic, or wire blend nodes and state machines into the root AnimGraph.
• Animation asset authoring. Creating or editing sequences, montages, blend spaces, notifies and curves is not covered by the current toolset.

Why the gap exists

Reflection exposes callable APIs and object properties, but reliable animation authoring also requires graph invariants, editor integration and validation. The native animation helpers cover specific inspection and repair tasks. A reflected class or property being visible does not turn it into a supported retargeting pipeline.

What the planned work has to cover

• Write primitives for state machines, so the agent can add and remove states and transitions rather than only describe them.
• AnimGraph authoring, so blend nodes and state machines can be wired into the root pose from a prompt.
• IK Rig and IK Retargeter asset creation and editing, including chain and bone mapping.
• Batch retargeting, so producing a retargeted animation set becomes one call instead of a manual pass in the editor.

Against the 5.7 generation