diff --git a/.specstory/history/2026-03-13_20-29-12Z-retargeting-focus-object-issue.md b/.specstory/history/2026-03-13_20-29-12Z-retargeting-focus-object-issue.md
index 3569336..e0b6a50 100644
--- a/.specstory/history/2026-03-13_20-29-12Z-retargeting-focus-object-issue.md
+++ b/.specstory/history/2026-03-13_20-29-12Z-retargeting-focus-object-issue.md
@@ -12,18 +12,18 @@ context:@.specstory/history/2026-03-10_17-24Z-version-and-commit-history-check.m
---
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
The user is reporting a bug where `bpy.ops.dlm.migrator_retarget_relations()` doesn't retarget `dof.focus_object` from orig to rep. Let me first understand the codebase structure and find the relevant code for this issue.
Let me start by reading the context files and exploring the codebase to understand how the migrator and retargeting works.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/.specstory/history/2026-03-13_20-28Z-camera-dof-focus-object-retargeting.md
@@ -36,7 +36,7 @@ _**Agent (kimi-k2.5)**_
Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/.specstory/history/2026-03-13_20-27Z-run-retarg-relatives-function-modification.md
@@ -49,7 +49,7 @@ _**Agent (kimi-k2.5)**_
@@ -80,12 +80,12 @@ Directory: **T:/Dynamic-Link-Manager** (12 files)
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
Now let me read the migrator.py file to understand the current state of the `run_retarg_relatives` function and see if the fix has been applied.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ops/migrator.py
@@ -98,7 +98,7 @@ _**Agent (kimi-k2.5)**_
Thought Process
Now I can see the issue. Looking at the `run_retarg_relatives` function (lines 461-491), I can see it handles:
@@ -125,11 +125,11 @@ The armature modifiers section ends at line 491:
I need to add the camera DOF code after this section. Let me add it right after line 491.Thought Process
The fix has been applied. Let me verify the change was made correctly by reading the file again.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ops/migrator.py
@@ -166,16 +166,16 @@ _**Agent (kimi-k2.5)**_
Thought Process
The fix has been applied correctly. The camera DOF focus_object retargeting code is now in place. Let me also check for any linter errors in the file.
@@ -213,12 +213,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
No linter errors. The fix has been successfully applied. Let me summarize what was done.Thought Process
The user wants a summary of the change I just made in 72 characters or less. This is the change I made to the migrator.py file where I added camera DOF focus_object retargeting.
@@ -260,7 +260,7 @@ I need to keep it under 72 characters. Let me check:
I'll go with something clear and concise.Thought Process
The user is saying that a character's eyes have a "Copy Location" constraint that targets the origchar, but the retargetrelatives function didn't retarget this constraint. Let me look at the `run_retarg_relatives` function to understand how it handles constraints and see what might be missing.
Let me read the relevant parts of the migrator.py file to understand the current constraint handling logic.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ops/migrator.py
@@ -296,7 +296,7 @@ _**Agent (kimi-k2.5)**_
Thought Process
The current code looks for objects that either:
@@ -318,13 +318,13 @@ The solution is to also scan ALL objects in the scene for constraints targeting
Let me modify the function to retarget constraints on ALL objects, not just the candidates.Thought Process
Good, I separated the constraint retargeting to scan all objects instead of just the candidates. This means that even objects in the original character's hierarchy (like eyes with Copy Location constraints) will have their constraints retargeted.
Let me verify the change and check for any linter errors.
@@ -384,12 +384,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
No linter errors. Let me verify the change looks correct.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ops/migrator.py
@@ -402,12 +402,12 @@ _**Agent (kimi-k2.5)**_
Thought Process
The fix looks good. The constraint retargeting is now separate from the candidates loop and scans all objects in the scene. This means that even objects in the original character's hierarchy (like eyes with Copy Location constraints) will have their constraints retargeted properly.Thought Process
The user is asking about an error in terminal 4.txt. Let me read that file to see what the error is.Tool use: **read_file_v2** • Read file: c:\Users\Nathan\.cursor\projects\t-Dynamic-Link-Manager\terminals\4.txt
@@ -439,7 +439,7 @@ _**Agent (kimi-k2.5)**_
Thought Process
The error is on line 104 of `tweak_tools.py`:
@@ -474,11 +474,11 @@ rep.pose.bones[name].bone.select_set(True)
Or we might need to use the select state on the bone differently. Let me fix this.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ops/tweak_tools.py
@@ -491,12 +491,12 @@ _**Agent (kimi-k2.5)**_
Thought Process
I see the issue. In Blender 4.x, `Bone.select` was removed. The correct API is `select_set(True)` or we need to use a different approach. Let me fix this.Thought Process
The fix has been applied. Let me verify there are no linter errors.
@@ -554,12 +554,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
The fix is complete. The error was caused by Blender 4.x API changes where `Bone.select` property was replaced with `select_set()` method.Thought Process
The user is asking about the terminal file again. Let me read the latest content to see if there are any new errors since my last fix.Tool use: **read_file_v2** • Read file: c:\Users\Nathan\.cursor\projects\t-Dynamic-Link-Manager\terminals\4.txt
@@ -589,7 +589,7 @@ _**Agent (kimi-k2.5)**_
Thought Process
The error shows that `bone.select_set(True)` also doesn't work. The issue is that in Blender 4.x, bone selection works differently. Let me check the correct API.
@@ -599,11 +599,11 @@ In Blender 4.x, you need to set the `select` attribute on the pose bone itself,
Let me fix this.Thought Process
Let me verify this fix and check for any linter errors.
@@ -661,12 +661,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
The fix looks good. The issue was that I was calling `select_set()` on the `bone` data object instead of the pose bone itself. In Blender 4.x, the `select_set()` method is on the PoseBone object, not the Bone object.Thought Process
The user is telling me this is Blender 5.0. Let me read the terminal file again to see if there are any new errors.Tool use: **run_terminal_command_v2** • Run command: tail -n 50 "c:\Users\Nathan\.cursor\projects\t-Dynamic-Link-Manager\terminals\4.txt"
@@ -709,12 +709,12 @@ Line |
Thought Process
Let me use PowerShell's Get-Content instead.Tool use: **run_terminal_command_v2** • Run command: Get-Content "c:\Users\Nathan\.cursor\projects\t-Dynamic-Link-Manager\terminals\4.txt" -Tail 50
@@ -784,7 +784,7 @@ AttributeError: 'PoseBone' object has no attribute 'select_set'
Thought Process
So in Blender 5.0, it seems `select_set()` doesn't exist on PoseBone either. Let me search online for the correct Blender 5.0 API for selecting bones in pose mode.
@@ -797,7 +797,7 @@ But actually, looking at the error, it seems like the API has changed in Blender
Actually, I should look at how other parts of the codebase handle bone selection, or use bpy.ops to select bones.
@@ -825,8 +825,6 @@ Result:
"rejected": false
}
```
-User decision: **accepted**
-
Status: **completed**
@@ -838,7 +836,7 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
Ah, so in Blender 5.0, the correct way to select a pose bone is:
@@ -853,11 +851,11 @@ rep.data.bones[name].select = True
Let me fix the code.Thought Process
Let me verify the fix and check for linter errors.
@@ -915,12 +913,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
Good, no linter errors. The fix is applied. In Blender 5.0, bone selection is done via `armature.data.bones[name].select = True` instead of the old `pose_bone.bone.select = True` or `pose_bone.select_set(True)` APIs.Thought Process
The user wants to modify the "tweak add both" feature to add ALL tweaks (every tweak bone), and they want me to use the Blender MCP to view their scene and identify all the tweak bones.
First, I need to check if there's a Blender MCP available. Let me look at the MCP servers available.
@@ -1016,12 +1014,12 @@ Directory: **C:\Users\Nathan\.cursor\projects\t-Dynamic-Link-Manager\mcps** (58
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
I can see there's a `user-blender` MCP with tools like `execute_blender_code` and `get_scene_info`. Let me read the tool schema to understand how to use it.Tool use: **read_file_v2** • Read file: C:\Users\Nathan\.cursor\projects\t-Dynamic-Link-Manager\mcps\user-blender\tools\execute_blender_code.json
@@ -1034,7 +1032,7 @@ _**Agent (kimi-k2.5)**_
Tool use: **read_file_v2** • Read file: C:\Users\Nathan\.cursor\projects\t-Dynamic-Link-Manager\mcps\user-blender\tools\get_scene_info.json
@@ -1047,12 +1045,12 @@ _**Agent (kimi-k2.5)**_
Thought Process
Now let me use the Blender MCP to get scene info and then execute code to find all tweak bones in the scene.
@@ -1097,7 +1095,7 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
I found 230 tweak bones across all armatures in the scene. Looking at the patterns:
@@ -1130,11 +1128,11 @@ I found 230 tweak bones across all armatures in the scene. Looking at the patter
Now let me look at the current tweak_tools.py to see how it currently works and then modify it to add ALL tweaks.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ops/tweak_tools.py
@@ -1147,7 +1145,7 @@ _**Agent (kimi-k2.5)**_
Thought Process
Now I understand the current implementation. The code has:
@@ -1163,11 +1161,11 @@ Let me modify the code so that when `limb='both'` (or maybe a new mode 'all'), i
Actually, the user said "tweak add both is add ALL" - so I'll modify the 'both' case to include ALL tweak bones found on the armature (any bone with "tweak" in the name).Thought Process
Now let me also update the docstrings and the bake function message to clarify that "both" now means "all tweaks". Also, I should update the module docstring.Thought Process
Let me verify the changes and check for linter errors.
@@ -1245,7 +1243,7 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
The changes are complete. Let me summarize what was changed:
@@ -1260,7 +1258,7 @@ This means:
The bones are filtered to only return those that actually exist on the armature, so it won't error if an armature doesn't have certain tweak bones.Thought Process
The user wants to create a new operator for body tweaks (no arm/leg). This includes spine tweaks like `spine_fk.003`, `spine_fk.002`, etc. They mentioned these are in the "torso tweaks" bone group in Rigify and are blue colored.
Let me first check the current operators.py to see how the tweak operators are defined, then add a new one for body tweaks.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ui/operators.py
@@ -1306,12 +1304,12 @@ _**Agent (kimi-k2.5)**_
Thought Process
Now let me look at the earlier operators to see how arm and leg tweaks are defined:
@@ -1359,7 +1357,7 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
Now I understand the structure. I need to:
@@ -1375,11 +1373,11 @@ The body tweaks should include:
Let me modify the code:Thought Process
Now I need to add the new body tweak operators to operators.py. I'll add them after the leg operators and before the "both" operators:Thought Process
Now I need to update the "remove both" and "bake both" labels/descriptions and add the new body operators to the OPERATOR_CLASSES list:Thought Process
Now update the bake both label/description:Thought Process
Now I need to add the new operators to the OPERATOR_CLASSES list:Thought Process
Now let me check the UI panels to see if there's a panel that displays the tweak operators, so I can add the new body operators there:
@@ -1543,12 +1541,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
Now I need to add the body tweak row to the UI panel:Thought Process
Now let me run the linter to check for errors:
@@ -1614,12 +1612,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
Everything is complete. Let me summarize the changes:Thought Process
Let me check the latest error in the terminal file to see what's happening with the tweak bake operations.Tool use: **run_terminal_command_v2** • Run command: Get-Content "c:\Users\Nathan\.cursor\projects\t-Dynamic-Link-Manager\terminals\4.txt" -Tail 60
@@ -1742,7 +1740,7 @@ C:\Program Files\Blender Foundation\Blender 5.0\5.0\scripts\startup\fa_panel_pan
Thought Process
The error is still happening. It seems my fix didn't get applied properly, or there's something else going on. The error is on line 116 of `tweak_tools.py`:
@@ -1753,11 +1751,11 @@ rep.data.bones[name].select = True
In Blender 5.0, `Bone` objects don't have a `select` attribute. Let me check the current state of the file to see what happened.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ops/tweak_tools.py
@@ -1770,12 +1768,12 @@ _**Agent (kimi-k2.5)**_
Thought Process
The file still has the old code. Let me check what's happening. Maybe my edit didn't save properly. Let me try again with the correct fix.Thought Process
Let me verify this change was applied and check for linter errors.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ops/tweak_tools.py
@@ -1808,7 +1806,7 @@ _**Agent (kimi-k2.5)**_
@@ -1846,12 +1844,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
The fix has been applied. In Blender 5.0, bone selection is done via `select_head` and `select_tail` properties rather than a single `select` property.Thought Process
The user wants me to check the Blender 5.0 API documentation. Let me search for the proper way to select bones in Blender 5.0.
@@ -1896,8 +1894,6 @@ Result:
"rejected": false
}
```
-User decision: **accepted**
-
Status: **completed**
@@ -1909,12 +1905,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
The search didn't give me the exact API for selecting bones in pose mode. Let me try a different search specifically for the select property or the new way to select bones in Blender 5.0.
@@ -1942,8 +1938,6 @@ Result:
"rejected": false
}
```
-User decision: **accepted**
-
Status: **completed**
@@ -1955,14 +1949,14 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
This is very helpful. It says that in Blender 5.0, bone selection state is now stored on the **pose bone** itself rather than on the bone data. So instead of `rep.data.bones[name].select_head`, we should use the pose bone directly.
Let me search for more specific information about how to select bones via the pose bone.
@@ -1984,8 +1978,6 @@ Result:
"url": "https://docs.blender.org/api/current/bpy.types.PoseBone.html"
}
```
-User decision: **accepted**
-
Status: **completed**
@@ -1997,12 +1989,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
Let me read this file to see the PoseBone API.Tool use: **read_file_v2** • Read file: C:\Users\Nathan\.cursor\projects\t-Dynamic-Link-Manager\agent-tools\2b59cbcc-6d6a-46a9-ac24-052eb68b5ed5.txt
@@ -2015,12 +2007,12 @@ _**Agent (kimi-k2.5)**_
Thought Process
Let me search for selection-related properties in this file.
@@ -2066,12 +2058,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
There's a `select` property on PoseBone at line 459. Let me read around that line.Tool use: **read_file_v2** • Read file: C:\Users\Nathan\.cursor\projects\t-Dynamic-Link-Manager\agent-tools\2b59cbcc-6d6a-46a9-ac24-052eb68b5ed5.txt
@@ -2084,7 +2076,7 @@ _**Agent (kimi-k2.5)**_
Thought Process
The `select` property exists on PoseBone in Blender 5.0. So the correct way to select bones in Blender 5.0 is:
@@ -2094,11 +2086,11 @@ rep.pose.bones[name].select = True
Let me update the tweak_tools.py file to use this.Thought Process
Let me verify the fix.
@@ -2156,12 +2148,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
Fixed. The correct API for Blender 5.0 is `rep.pose.bones[name].select = True`.Thought Process
Let me look at the current constraint retargeting logic in `run_retarg_relatives`. The user is saying that constraints on bones of characters that aren't the origchar or repchar are not being remapped.
Let me read the migrator.py file to see the current implementation.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ops/migrator.py
@@ -2199,7 +2191,7 @@ _**Agent (kimi-k2.5)**_
Thought Process
I see the issue. The code at lines 490-494 handles object-level constraints (where `c.target` is an object), but it doesn't handle **bone constraints** on other characters.
@@ -2208,11 +2200,11 @@ Bone constraints are on `ob.pose.bones[...].constraints`, not on `ob.constraints
I need to add logic to iterate through all armatures and check their pose bones' constraints, retargeting any that point to `orig` to point to `rep` instead.Thought Process
Good, I've added bone constraint retargeting. Now any constraint on any bone of any armature that targets the original character will be retargeted to the replacement character.
@@ -2270,12 +2262,12 @@ Status: **completed**
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
Good, no linter errors. The fix adds a new loop that goes through all armatures in the scene and checks each pose bone's constraints, retargeting any that point to `orig` to point to `rep` instead.Thought Process
The user wants me to bump the version to 0.3.0 and update the changelog. Let me first find the version file and the changelog to understand the current format.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/blender_manifest.toml
-
-
-Tool use: **read_file_v2**
-
-
-
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
@@ -2332,7 +2311,20 @@ Directory: **T:/Dynamic-Link-Manager** (1 file)
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
+
+
+Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/blender_manifest.toml
+
+
+Tool use: **read_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
@@ -2348,7 +2340,7 @@ Directory: **T:/Dynamic-Link-Manager** (0 files)
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
@@ -2371,12 +2363,12 @@ Directory: **T:/Dynamic-Link-Manager** (4 files)
-_**Agent (kimi-k2.5)**_
+_**Agent (composer-2-fast)**_
Thought Process
Good, I found the version in `blender_manifest.toml` and there's a `CHANGELOG.md`. Let me read the changelog to see the format.Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/CHANGELOG.md
@@ -2389,7 +2381,7 @@ _**Agent (kimi-k2.5)**_
Thought Process
Now I need to:
@@ -2405,7 +2397,7 @@ Let me summarize the changes made in this session:
6. Fixed Blender 5.0 bone selection API (pose_bone.select)
7. Renamed UI labels from "Arm & Leg" to "All" for the both modeThought Process
Done. I've bumped the version to 0.3.0 and updated the changelog with the key changes from this session.Thought Process
+**Planning feature set**
+
+I'm planning the mutually exclusive selector and ARP rig behavior adjustments.
+Tool use: **ripgrep_raw_search**
+
+
+Parameters:
+
+```json
+{
+ "caseInsensitive": true,
+ "glob": "*.py",
+ "path": "T:/Dynamic-Link-Manager",
+ "pattern": "CharMig|character.*mig|migrator"
+}
+```
+Additional data:
+
+```json
+{
+ "isPruned": true,
+ "outputMode": "content",
+ "path": "T:/Dynamic-Link-Manager",
+ "pattern": "CharMig|character.*mig|migrator",
+ "topFiles": [
+ {
+ "matchCount": 62,
+ "uri": "ui\\operators.py"
+ },
+ {
+ "matchCount": 62,
+ "uri": "ui/operators.py"
+ },
+ {
+ "matchCount": 2,
+ "uri": "ui/properties.py"
+ },
+ {
+ "matchCount": 13,
+ "uri": "ui/panels.py"
+ },
+ {
+ "matchCount": 3,
+ "uri": "ops/migrator.py"
+ },
+ {
+ "matchCount": 1,
+ "uri": "ops/__init__.py"
+ }
+ ],
+ "totalFiles": 6,
+ "totalMatches": 143
+}
+```
+Status: **completed**
+
+
+
+
+
+
+Tool use: **ripgrep_raw_search**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+
+Tool use: **ripgrep_raw_search**
+
+
+Parameters:
+
+```json
+{
+ "caseInsensitive": false,
+ "glob": "*.py",
+ "path": "T:/Dynamic-Link-Manager",
+ "pattern": "remove_original|Remove Original"
+}
+```
+Additional data:
+
+```json
+{
+ "isPruned": true,
+ "outputMode": "content",
+ "path": "T:/Dynamic-Link-Manager",
+ "pattern": "remove_original|Remove Original",
+ "topFiles": [
+ {
+ "matchCount": 1,
+ "uri": "ui/panels.py"
+ },
+ {
+ "matchCount": 4,
+ "uri": "ui/operators.py"
+ },
+ {
+ "matchCount": 4,
+ "uri": "ui\\operators.py"
+ }
+ ],
+ "totalFiles": 3,
+ "totalMatches": 9
+}
+```
+Status: **completed**
+
+
+
+
+
+
+Tool use: **ripgrep_raw_search**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ui/operators.py
+
+
+Tool use: **read_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ui/properties.py
+
+
+Tool use: **read_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ui/panels.py
+
+
+Tool use: **read_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/utils/__init__.py
+
+
+Tool use: **read_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+Thought Process
+**Clarifying rigify vs ARP**
+
+I'm implementing a mutually exclusive selector for Rigify vs ARP enum at the top of CharMig.
+
+**Handling linked collection cleanup**
+
+I'm walking the hierarchy to find the correct collection for removal in linked collections. I'm using Blender MCP to read scene context and identify rigify vs arp implementations. Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ops/fk_rotations.py
+
+
+Tool use: **read_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+Thought Process
+**Adapting for ARP**
+
+I'm adapting Rigify-specific patterns for ARP, focusing on parallel migration steps and collection hierarchy handling.
+
+**Planning arp behavior**
+
+I'm implementing a modular approach with `rig_profile` enum and dispatch table for ARP behavior.
+
+**Tracing armature collections**
+
+I'm tracing the armature collection hierarchy to identify library instances for proper cleanup.
+Tool use: **mcp-blender-user-blender-execute_blender_code**
+
+
+Parameters:
+
+```json
+{
+ "tools": [
+ {
+ "name": "execute_blender_code",
+ "parameters": "{\"code\":\"import bpy\\nsel = [o.name for o in bpy.context.selected_objects]\\nact = bpy.context.active_object.name if bpy.context.active_object else None\\nprint('ACTIVE:', act)\\nprint('SELECTED:', sel)\\nfor o in bpy.context.selected_objects:\\n if o.type=='ARMATURE':\\n print('ARMATURE', o.name, 'lib=', getattr(o,'library',None), 'override=', getattr(getattr(o,'override_library',None),'reference',None))\\n uc = list(getattr(o,'users_collection',[]) or [])\\n print(' users_collection:', [c.name for c in uc])\\n for c in uc[:5]:\\n print(' coll', c.name, 'lib=', getattr(c,'library',None))\",\"user_prompt\":\"View scene selection for CharMig planning\"}",
+ "serverName": "blender"
+ }
+ ]
+}
+```
+Additional data:
+
+```json
+{
+ "status": "success"
+}
+```
+Result:
+
+```json
+{
+ "result": "{\"content\":[{\"type\":\"text\",\"text\":\"Code executed successfully: ACTIVE: rig.001\\nSELECTED: []\\n\"}],\"structuredContent\":{\"result\":\"Code executed successfully: ACTIVE: rig.001\\nSELECTED: []\\n\"}}"
+}
+```
+Status: **completed**
+
+
+
+
+
+
+Tool use: **mcp-blender-user-blender-execute_blender_code**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+MCP shows no selected objects in the connected Blender session (`ACTIVE: rig.001`, `SELECTED: []`). Gathering implementation details: your message was truncated after “modify it according to”.
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+
+Tool use: **ask_question**
+
+
+Parameters:
+
+```json
+{
+ "questions": [
+ {
+ "id": "arp_steps",
+ "options": [
+ {
+ "id": "remove_only",
+ "label": "Remove Original + collection fix first; ARP bone lists later"
+ },
+ {
+ "id": "core_mig",
+ "label": "All migrator steps (CopyAttr, MigNLA, CustProps, BoneConst, Retarg, BBody, FKRot, Tweaks) with ARP tables where needed"
+ },
+ {
+ "id": "minimal",
+ "label": "UI enum + wiring only; stub ARP to Rigify until we iterate"
+ }
+ ],
+ "prompt": "Which CharMig steps should branch on Rigify vs ARP first? (We can add the enum everywhere but implement ARP logic incrementally.)"
+ }
+ ],
+ "title": "ARP migration scope"
+}
+```
+Additional data:
+
+```json
+{
+ "currentSelections": {
+ "arp_steps": [
+ "remove_only"
+ ]
+ },
+ "freeformTexts": {
+ "arp_steps": ""
+ },
+ "status": "submitted"
+}
+```
+Result:
+
+```json
+{
+ "answers": [
+ {
+ "freeformText": "",
+ "questionId": "arp_steps",
+ "selectedOptionIds": [
+ "remove_only"
+ ]
+ }
+ ]
+}
+```
+User decision: **accepted**
+
+Status: **completed**
+
+
+
+
+
+
+Tool use: **ask_question**
+
+
+
+_**Agent (composer-2-fast)**_
+
+Thought Process
+**Planning migration**
+
+I'm prioritizing the Remove Original + collection fix first, deferring ARP bone lists for later. I'm deferring ARP bone structure migrations to later while keeping the selector in place for remove-only scope.
+
+**Designing collection resolution**
+
+I'm implementing the collection resolution algorithm to identify character root collections by walking up the hierarchy.
+
+**Mapping collection hierarchy**
+
+I'm implementing collection hierarchy mapping using bpy.data.collections and children links to trace parents. I'm fixing the collection mapping to find the direct child collection of the linked library root instead of just picking the first matching name.
+
+**Fixing collection deletion**
+
+I'm correcting the deletion logic to target the root linked collection instead of a nested one.
+
+**Tracing orig**
+
+I'm tracing the orig reference to understand the linked collection structure.
+
+**Checking library overrides**
+
+I'm checking override_library on objects to handle armature library references properly.
+
+**Creating migration plan**
+
+I'm implementing a unified resolver for collection removal with a new enum property for rig type identification.
+Tool use: **create_plan**
+
+
+Parameters:
+
+```json
+{
+ "name": "CharMig Rigify/ARP + Remove Original",
+ "overview": "Add a Rigify vs ARP enum at the top of the Character Migrator panel, wire it into scene properties for future branching, and implement a robust linked/local collection resolution path for Remove Original (priority: ARP-style linked layouts and overrides). Full ARP bone migration tables are deferred per your preference.",
+ "plan": "# CharMig rig selector and Remove Original fixes\n\n## Context (from codebase)\n\n- CharMig UI lives in [`ui/panels.py`](ui/panels.py) (box labeled \"Character Migrator\").\n- Scene props are [`DynamicLinkManagerProperties`](ui/properties.py) in [`ui/properties.py`](ui/properties.py).\n- Remove Original is [`DLM_OT_migrator_remove_original`](ui/operators.py) (~lines 492–575); it uses [`collection_containing_armature`](utils/__init__.py) then `bpy.data.collections.remove(coll)`, else `bpy.data.objects.remove(orig)`.\n- [`collection_containing_armature`](utils/__init__.py) picks one of `armature.users_collection` using a **Rigify-oriented** name heuristic (`_Rigify`, base name match) and falls back to `colls[0]` — fragile for nested linked collections and non-Rigify names (ARP).\n\nBlender MCP in this session reported no selected objects (`ACTIVE: rig.001`, `SELECTED: []`), so collection layout was not visually verified; implementation should be driven by the resolution algorithm below and your real files.\n\n## 1. Mutually exclusive Rigify / ARP selector (top of CharMig)\n\n- Add an **`EnumProperty`** on `DynamicLinkManagerProperties`, e.g. `migrator_rig_family` with items `RIGIFY` and `ARP` (single choice = mutually exclusive).\n- In [`ui/panels.py`](ui/panels.py), **first row inside the Character Migrator box** (before \"Automatic pair discovery\"): draw this enum (e.g. `row.prop(props, \"migrator_rig_family\", expand=True)` or labeled dropdown).\n- Register/update [`ui/__init__.py`](ui/__init__.py) if needed so the property is saved with the scene.\n- **Scope for this phase:** operators other than Remove Original can **read** `props.migrator_rig_family` but need no behavior change yet (future ARP bone tables).\n\n## 2. Remove Original: unified collection resolver (ARP-first behavior, Rigify benefits)\n\nExtract logic from the operator into a small helper module (recommended: **`utils/remove_original.py`** or **`utils/collection_resolve.py`**) with a single entry point, e.g. `resolve_collection_for_remove_original(orig, rig_family) -\u003e bpy.types.Collection | None`, documented with assumptions.\n\n**Goals (aligned with your description):**\n\n- Consider **all** `orig.users_collection` entries, not only the first heuristic match.\n- **Walk the collection hierarchy upward** (find parent collections via scanning `bpy.data.collections` for `coll in parent.children`) to avoid picking a **nested** linked child when the intent is to remove the **top-level linked instance** (or the correct subtree).\n- **Linked vs local:** use `Collection.library` / `Collection.override_library` and `Object.library` / `Object.override_library` to distinguish:\n - Linked collection hierarchy,\n - **Locally overridden** armature/collection (still sitting under a linked tree),\n - Fully local objects.\n- **Selection rule (concrete approach):**\n - Build candidate collections that contain `orig`.\n - Score or filter candidates using: linked root (no parent in scene that is also a user of the same library), minimal enclosing set for the character, or **highest ancestor** that still “owns” the linked asset (exact tie-break documented in code comments).\n - If **no safe collection** can be removed (e.g. linked data cannot be removed), **fall back** to `bpy.data.objects.remove(orig, do_unlink=True)` (and optionally unlink from collections only) instead of deleting the wrong nested collection.\n\n**ARP-specific:** same resolver, but **do not** rely on `_Rigify` string stripping in [`collection_containing_armature`](utils/__init__.py) when `rig_family == ARP` — use neutral base name from `orig.name` (strip common suffixes like `.001` only) or prefer hierarchy-linked resolution over name matching.\n\n**Rigify:** keep backward compatibility by either calling the new resolver for both, or delegating old behavior only if you need a safety flag; prefer **one path** to reduce bugs.\n\nUpdate [`DLM_OT_migrator_remove_original`](ui/operators.py) to:\n\n1. Read `context.scene.dynamic_link_manager.migrator_rig_family`.\n2. Call the resolver; remove that collection if returned; else existing object fallback.\n3. Preserve current action purge + rep action rename behavior unchanged.\n\n## 3. Deferred (explicitly out of this phase)\n\n- ARP-specific bone naming for MigFKRot, tweak lists, `get_pair_automatic`, base-body mesh heuristics — **not implemented** until you add a follow-up task; the enum is the hook.\n\n## Files to touch\n\n| File | Change |\n|------|--------|\n| [`ui/properties.py`](ui/properties.py) | `EnumProperty` `migrator_rig_family` |\n| [`ui/panels.py`](ui/panels.py) | Draw enum at top of CharMig box |\n| New `utils/...py` | Collection resolution + optional small tests via docstring examples |\n| [`ui/operators.py`](ui/operators.py) | Remove Original uses resolver + `rig_family` |\n| [`utils/__init__.py`](utils/__init__.py) | Optionally re-export helper or leave `collection_containing_armature` for other callers (MigBBody) — only change if we need a shared neutral “base name” helper |\n\n## Risk / validation\n\n- Removing linked collections can **fail or be disallowed** by Blender; the resolver must **never** assume `bpy.data.collections.remove` always succeeds — catch, report, fallback to object removal.\n- Test manually: linked char at root, linked char nested one level, local override of armature in linked collection, Rigify and ARP-named rigs.\n",
+ "todos": [
+ {
+ "content": "Add migrator_rig_family EnumProperty and draw at top of CharMig box in panels.py",
+ "id": "enum-prop-ui",
+ "status": "pending"
+ },
+ {
+ "content": "Implement resolve_collection_for_remove_original(orig, rig_family) with linked/override hierarchy walk + safe fallback",
+ "id": "collection-resolver",
+ "status": "pending"
+ },
+ {
+ "content": "Refactor DLM_OT_migrator_remove_original to use resolver; keep action purge/rename logic",
+ "id": "wire-remove-original",
+ "status": "pending"
+ }
+ ]
+}
+```
+Additional data:
+
+```json
+{
+ "hasOpenedEditor": true,
+ "planUri": "file:///c%3A/Users/Nathan/.cursor/plans/charmig_rigify_arp_%2B_remove_original_f7b9b429.plan.md",
+ "reviewData": {
+ "isShowingInput": false,
+ "selectedOption": "none",
+ "status": "Requested"
+ }
+}
+```
+Result:
+
+```json
+{
+ "rejected": {}
+}
+```
+Status: **completed**
+
+
+
+
+
+
+Tool use: **create_plan**
+
+
+
+---
+
+_**User**_
+
+Implement the plan as specified, it is attached for your reference. Do NOT edit the plan file itself.
+
+To-do's from the plan have already been created. Do not create them again. Mark them as in_progress as you work, starting with the first one. Don't stop until you have completed all the to-dos.
+
+---
+
+_**Agent (composer-2-fast)**_
+
+
+Implementing the plan: starting with the enum property and UI, then the collection resolver, then wiring Remove Original.
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+
+Tool use: **todo_write**
+
+
+Parameters:
+
+```json
+{
+ "merge": true
+}
+```
+Result:
+
+```json
+{
+ "finalTodos": [
+ {
+ "content": "Add migrator_rig_family EnumProperty and draw at top of CharMig box in panels.py",
+ "id": "enum-prop-ui",
+ "status": "in_progress"
+ },
+ {
+ "content": "Implement resolve_collection_for_remove_original(orig, rig_family) with linked/override hierarchy walk + safe fallback",
+ "id": "collection-resolver",
+ "status": "pending"
+ },
+ {
+ "content": "Refactor DLM_OT_migrator_remove_original to use resolver; keep action purge/rename logic",
+ "id": "wire-remove-original",
+ "status": "pending"
+ }
+ ],
+ "initialTodos": [
+ {
+ "content": "Add migrator_rig_family EnumProperty and draw at top of CharMig box in panels.py",
+ "id": "enum-prop-ui",
+ "status": "pending"
+ },
+ {
+ "content": "Implement resolve_collection_for_remove_original(orig, rig_family) with linked/override hierarchy walk + safe fallback",
+ "id": "collection-resolver",
+ "status": "pending"
+ },
+ {
+ "content": "Refactor DLM_OT_migrator_remove_original to use resolver; keep action purge/rename logic",
+ "id": "wire-remove-original",
+ "status": "pending"
+ }
+ ],
+ "success": true,
+ "wasMerge": true
+}
+```
+Status: **completed**
+
+
+
+
+
+
+Tool use: **todo_write**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ui/properties.py
+
+
+Tool use: **read_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ui/panels.py
+
+
+Tool use: **read_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+Tool use: **read_file_v2** • Read file: T:/Dynamic-Link-Manager/ui/__init__.py
+
+
+Tool use: **read_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+Thought Process
+**Updating migrator panel**
+
+I'm adding EnumProperty for the character migrator mode. I'm implementing collection parent lookup using `bpy.data.collections` iteration with `coll in c.children` check. I'm fixing the collection parent lookup to use `coll.name in parent.children` check with proper Blender 3.x+ API. I'm walking up the collection hierarchy to find linked roots using `bpy.types.Collection.library`. I'm traversing collection hierarchies to identify linked roots using `bpy.types.Collection.library` and `override_library`.
+
+**Traversing collection hierarchies**
+
+I'm identifying linked roots by checking `c.library is not None` and tracking parent collections with the same library.
+
+**Identifying instance roots**
+
+I'm finding the ancestor collection that serves as the instance root for deletion instead of the subcollection.
+
+**Finding instance roots**
+
+I'm identifying the outermost collection containing the armature by selecting the minimum depth collection in the hierarchy. I'm computing the depth of each candidate collection by walking up from each to find the minimum ancestor collection.
+
+I'm identifying the outermost collection by finding the one with the minimum depth from the scene root. I'm determining the outermost collection by finding the one with the smallest depth in the tree hierarchy.
+
+**Analyzing collection linking**
+
+I'm realizing that in Blender's linking model, objects may only register as being in the top-level collection, not necessarily their nested sub-collections. I'm implementing algorithm to walk up collection hierarchy and prefer removing linked root collections. I found that Blender objects are explicitly linked to collections, so parent collections don't necessarily contain objects, and removing a collection may not automatically remove child objects.
+
+**Finding linked root**
+
+I'm walking up parent collections to find the linked root for the character. I'm traversing parent chains to identify the topmost safe-to-remove ancestor for the linked root.
+
+**Tracing collection hierarchy**
+
+I'm verifying Blender's collection hierarchy rules to ensure correct parent-child relationships. I'm implementing `parent_of()` to find collection parents by scanning `scene.collection.children`. I'm identifying the topmost collection in the chain by walking up from the user collection to the scene root. I'm determining the outermost collection in the hierarchy by traversing up from the user collection to the scene root, accounting for potential linking behavior in Blender. I'm identifying the target collection as the outermost ancestor of the shallowest-depth user collection to properly remove linked collections.
+
+I'm implementing the logic to identify the target collection by finding the outermost ancestor of the shallowest user collection. I'm clarifying the logic for handling linked collections and local overrides during hierarchy traversal.
+
+I'm implementing a collection hierarchy resolver using Blender's override_library API. I'm walking up the collection chain to find the outermost parent.
+
+I found that `chain[-1]` is the collection directly under Scene Collection's subtree root, not necessarily its direct child. I'm clarifying the relationship between `scene.collection` and `bpy.data.collections` to understand the collection tree structure. I'm clarifying the collection hierarchy API: using `bpy.data.collections` to find parents instead of `collection.users_collection` which is for object relationships. I'm implementing a collection parent lookup using iterative traversal of all collections.
+
+I'm implementing iterative parent collection lookup with `bpy.data.collections` traversal. I'm identifying the outermost collection in the chain by checking if the parent is scene.collection. I found that `bpy.context.scene.collection` is the master collection and all collections are children of it. Tool use: **edit_file_v2** • Edit file: T:/Dynamic-Link-Manager/ui/properties.py
+
+
+Tool use: **edit_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+
+Tool use: **edit_file_v2** • Edit file: T:/Dynamic-Link-Manager/ui/properties.py
+
+
+Tool use: **edit_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+
+Tool use: **edit_file_v2** • Edit file: T:/Dynamic-Link-Manager/ui/panels.py
+
+
+Tool use: **edit_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+
+Tool use: **todo_write**
+
+
+Parameters:
+
+```json
+{
+ "merge": true
+}
+```
+Result:
+
+```json
+{
+ "finalTodos": [
+ {
+ "content": "Add migrator_rig_family EnumProperty and draw at top of CharMig box in panels.py",
+ "id": "enum-prop-ui",
+ "status": "completed"
+ },
+ {
+ "content": "Implement resolve_collection_for_remove_original(orig, rig_family) with linked/override hierarchy walk + safe fallback",
+ "id": "collection-resolver",
+ "status": "in_progress"
+ },
+ {
+ "content": "Refactor DLM_OT_migrator_remove_original to use resolver; keep action purge/rename logic",
+ "id": "wire-remove-original",
+ "status": "pending"
+ }
+ ],
+ "initialTodos": [
+ {
+ "content": "Add migrator_rig_family EnumProperty and draw at top of CharMig box in panels.py",
+ "id": "enum-prop-ui",
+ "status": "in_progress"
+ },
+ {
+ "content": "Implement resolve_collection_for_remove_original(orig, rig_family) with linked/override hierarchy walk + safe fallback",
+ "id": "collection-resolver",
+ "status": "pending"
+ },
+ {
+ "content": "Refactor DLM_OT_migrator_remove_original to use resolver; keep action purge/rename logic",
+ "id": "wire-remove-original",
+ "status": "pending"
+ }
+ ],
+ "success": true,
+ "wasMerge": true
+}
+```
+Status: **completed**
+
+
+
+
+
+
+Tool use: **todo_write**
+
+
+
+_**Agent (composer-2-fast)**_
+
+
+
+Tool use: **edit_file_v2** • Edit file: T:/Dynamic-Link-Manager/utils/remove_original.py
+
+
+Tool use: **edit_file_v2**
+
+
+
+_**Agent (composer-2-fast)**_
+
+Thought Process
+**Checking Blender API**
+
+I'm verifying if `children_recursive` exists on `bpy.types.Collection` in Blender 5. I'm using `_parent_collection` to walk up from collection to scene root and count depth.
+
+**Verifying collection hierarchy**
+
+I'm checking if collections can appear in multiple parents in Blender's collection hierarchy.
+
+**Refining collection traversal**
+
+I'm removing `children_recursive` and simplifying the chain walk to use walk-up only for depth, while fixing the `_pick_remove_target_from_chain` logic for linked origins and collections. I'm simplifying `_pick_remove_target_from_chain` to default to `chain[-1]` for the topmost direct child of scene.collection. I'm simplifying `_pick_remove_target_from_chain` to always default to `chain[-1]` for the main fix. Tool use: **edit_file_v2** • Edit file: T:/Dynamic-Link-Manager/utils/remove_original.py
+
+
+Tool use: **edit_file_v2**
+
+
diff --git a/.specstory/statistics.json b/.specstory/statistics.json
index 8e284c8..a2d88e4 100644
--- a/.specstory/statistics.json
+++ b/.specstory/statistics.json
@@ -7,7 +7,7 @@
"end_timestamp": "2026-02-18T16:39:18-07:00",
"markdown_size_bytes": 789326,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"14305e53-b999-4885-8511-924b3957eb04": {
"user_message_count": 1,
@@ -16,7 +16,7 @@
"end_timestamp": "2026-03-13T14:21:38-06:00",
"markdown_size_bytes": 659,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:47:32Z"
+ "last_updated": "2026-03-25T20:24:24Z"
},
"15ca8ebd-f444-47c7-8df3-efe4c9e8386f": {
"user_message_count": 1,
@@ -25,7 +25,7 @@
"end_timestamp": "2026-03-10T10:07:27-06:00",
"markdown_size_bytes": 9952,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"315fde62-50dd-4c50-9d07-4089c4952f9d": {
"user_message_count": 1,
@@ -34,7 +34,7 @@
"end_timestamp": "2026-03-13T14:28:15-06:00",
"markdown_size_bytes": 1169,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:47:32Z"
+ "last_updated": "2026-03-25T20:24:24Z"
},
"3178f110-1a60-43b6-9d92-e189b2342c51": {
"user_message_count": 1,
@@ -43,7 +43,7 @@
"end_timestamp": "2026-03-13T14:27:44-06:00",
"markdown_size_bytes": 880,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:47:32Z"
+ "last_updated": "2026-03-25T20:24:24Z"
},
"3d77131c-052c-4be7-b2e4-2326abcd6957": {
"user_message_count": 1,
@@ -52,7 +52,7 @@
"end_timestamp": "2026-03-10T10:15:27-06:00",
"markdown_size_bytes": 8570,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"3f042eb5-f457-4e06-9b56-ee14ac7afd90": {
"user_message_count": 1,
@@ -61,7 +61,7 @@
"end_timestamp": "2026-02-19T11:53:38-07:00",
"markdown_size_bytes": 36111,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"49972a6d-d81d-4647-b6b5-9c92f8599474": {
"user_message_count": 1,
@@ -70,7 +70,7 @@
"end_timestamp": "2026-03-10T10:09:58-06:00",
"markdown_size_bytes": 12876,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"4ce1e38a-1e85-411b-93ee-a378eef56146": {
"user_message_count": 91,
@@ -79,7 +79,7 @@
"end_timestamp": "2025-08-20T16:43:14-06:00",
"markdown_size_bytes": 784540,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"4f67a3b1-1537-4680-9763-d0c401c58d7e": {
"user_message_count": 1,
@@ -88,7 +88,7 @@
"end_timestamp": "2026-03-10T10:19:09-06:00",
"markdown_size_bytes": 71838,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"577872ce-9067-4be5-bdb0-89d48d9424ca": {
"user_message_count": 1,
@@ -97,7 +97,7 @@
"end_timestamp": "2026-03-10T10:25:51-06:00",
"markdown_size_bytes": 75382,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"73bda206-5380-40cc-afb8-b5fc426fc0dc": {
"user_message_count": 1,
@@ -106,7 +106,7 @@
"end_timestamp": "2026-03-10T10:17:09-06:00",
"markdown_size_bytes": 11226,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"788ea0b5-9ff6-440d-993d-a05234d95c5a": {
"user_message_count": 1,
@@ -115,7 +115,7 @@
"end_timestamp": "2026-03-10T10:22:17-06:00",
"markdown_size_bytes": 69439,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"7ac38efd-6dd0-42da-9a01-f7d5062046e9": {
"user_message_count": 1,
@@ -124,7 +124,7 @@
"end_timestamp": "2026-03-10T10:23:09-06:00",
"markdown_size_bytes": 61512,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"7d4946a5-6831-4e4c-90f7-8e9a96fa62d5": {
"user_message_count": 1,
@@ -133,7 +133,7 @@
"end_timestamp": "2026-03-10T11:03:21-06:00",
"markdown_size_bytes": 107701,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"986c4b32-627a-4d70-9edb-fcaf30e33ee2": {
"user_message_count": 1,
@@ -142,7 +142,7 @@
"end_timestamp": "2026-03-10T10:04:31-06:00",
"markdown_size_bytes": 9282,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"9d47ad7e-a8c1-4b72-b7cb-d27d2232bf50": {
"user_message_count": 1,
@@ -151,7 +151,7 @@
"end_timestamp": "2026-03-10T11:24:38-06:00",
"markdown_size_bytes": 13504,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"9dca74c5-e8b6-4191-ae49-949e8c742e4d": {
"user_message_count": 1,
@@ -160,7 +160,7 @@
"end_timestamp": "2026-03-10T10:07:46-06:00",
"markdown_size_bytes": 58095,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"a6a5c4fa-5957-4577-9323-13723784a540": {
"user_message_count": 1,
@@ -169,7 +169,7 @@
"end_timestamp": "2026-03-10T10:05:03-06:00",
"markdown_size_bytes": 1116,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"b450a89e-cc97-4ef3-a775-fe26c81c7d0b": {
"user_message_count": 1,
@@ -178,7 +178,7 @@
"end_timestamp": "2026-03-10T10:09:41-06:00",
"markdown_size_bytes": 3758,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"be47994a-f179-482a-9af2-4d208a01d324": {
"user_message_count": 1,
@@ -187,7 +187,7 @@
"end_timestamp": "2026-03-10T10:07:14-06:00",
"markdown_size_bytes": 2904,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"cb944af7-6783-4830-acd6-af18544ff67e": {
"user_message_count": 1,
@@ -196,7 +196,7 @@
"end_timestamp": "2026-03-10T10:10:35-06:00",
"markdown_size_bytes": 5188,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"d613f220-0255-4158-8a5d-4f93d50aa47b": {
"user_message_count": 1,
@@ -205,7 +205,7 @@
"end_timestamp": "2026-03-10T10:17:43-06:00",
"markdown_size_bytes": 3856,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"daaa74ba-72d7-4686-a95b-ae6379d17ee9": {
"user_message_count": 1,
@@ -214,7 +214,7 @@
"end_timestamp": "2026-03-10T10:25:21-06:00",
"markdown_size_bytes": 3302,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"dd526f99-b20d-401d-abac-ea6770d3a424": {
"user_message_count": 1,
@@ -223,16 +223,16 @@
"end_timestamp": "2026-03-10T11:25:10-06:00",
"markdown_size_bytes": 24412,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
},
"f09112e7-3ddb-4e4e-b91a-418e38b881bf": {
- "user_message_count": 12,
- "agent_message_count": 162,
+ "user_message_count": 14,
+ "agent_message_count": 199,
"start_timestamp": "2026-03-13T14:29:12-06:00",
"end_timestamp": "2026-03-13T14:29:12-06:00",
- "markdown_size_bytes": 199500,
+ "markdown_size_bytes": 230631,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:28:01Z"
},
"f1bb25c5-e5d4-4aad-88a3-f562ab877040": {
"user_message_count": 1,
@@ -241,7 +241,7 @@
"end_timestamp": "2026-03-10T10:08:58-06:00",
"markdown_size_bytes": 21001,
"provider": "cursoride",
- "last_updated": "2026-03-25T19:45:41Z"
+ "last_updated": "2026-03-25T20:22:34Z"
}
}
}
diff --git a/ui/operators.py b/ui/operators.py
index b0664a5..7126ef8 100644
--- a/ui/operators.py
+++ b/ui/operators.py
@@ -9,7 +9,7 @@ from bpy.types import Operator
from bpy.props import StringProperty, BoolProperty
from bpy.props import StringProperty, IntProperty
-from ..utils import collection_containing_armature
+from ..utils.remove_original import resolve_collection_for_remove_original
ADDON_NAME = __package__.rsplit(".", 1)[0] if "." in __package__ else __package__
@@ -531,9 +531,10 @@ class DLM_OT_migrator_remove_original(Operator):
if removed_actions:
self.report({"INFO"}, f"Removed {len(removed_actions)} action(s) from original")
+ props = context.scene.dynamic_link_manager
+ rig_family = getattr(props, "migrator_rig_family", "RIGIFY")
try:
- # Try to find and delete the collection containing the original character
- coll = collection_containing_armature(orig)
+ coll = resolve_collection_for_remove_original(orig, rig_family, context.scene)
if coll:
coll_name = coll.name # Store name BEFORE removal (RNA invalidates after remove)
context.scene.dynamic_link_manager.original_character = None
@@ -541,10 +542,14 @@ class DLM_OT_migrator_remove_original(Operator):
bpy.data.collections.remove(coll)
self.report({"INFO"}, f"Removed collection: {coll_name}")
except Exception as remove_err:
- # Collection may have already been removed by another process
self.report({"WARNING"}, f"Collection {coll_name} removal issue: {remove_err}")
+ try:
+ bpy.data.objects.remove(orig, do_unlink=True)
+ self.report({"INFO"}, f"Removed original object: {name}")
+ except Exception as e2:
+ self.report({"ERROR"}, f"Could not remove original after collection failure: {e2}")
+ return {"CANCELLED"}
else:
- # Fallback: just delete the armature object
bpy.data.objects.remove(orig, do_unlink=True)
context.scene.dynamic_link_manager.original_character = None
self.report({"INFO"}, f"Removed original character: {name}")
diff --git a/ui/panels.py b/ui/panels.py
index 5b14a67..2923f45 100644
--- a/ui/panels.py
+++ b/ui/panels.py
@@ -76,6 +76,8 @@ class DLM_PT_main_panel(Panel):
box = layout.box()
box.label(text="Character Migrator")
row = box.row()
+ row.prop(props, "migrator_rig_family", expand=True)
+ row = box.row()
row.prop(props, "migrator_mode", text="Automatic pair discovery")
row = box.row()
row.prop(props, "original_character", text="Original")
diff --git a/ui/properties.py b/ui/properties.py
index a7bce05..4683ca0 100644
--- a/ui/properties.py
+++ b/ui/properties.py
@@ -5,7 +5,14 @@
import bpy
from bpy.types import PropertyGroup
-from bpy.props import IntProperty, StringProperty, BoolProperty, CollectionProperty, PointerProperty
+from bpy.props import (
+ IntProperty,
+ StringProperty,
+ BoolProperty,
+ CollectionProperty,
+ PointerProperty,
+ EnumProperty,
+)
class SearchPathItem(PropertyGroup):
@@ -41,6 +48,17 @@ class DynamicLinkManagerProperties(PropertyGroup):
)
selected_asset_path: StringProperty(name="Selected Asset Path", default="")
+ # Character migrator: rig family (future ARP-specific migration; Remove Original uses it for collection resolution)
+ migrator_rig_family: EnumProperty(
+ name="Rig family",
+ description="Rigify vs Auto-Rig Pro: affects Remove Original collection detection and future bone heuristics",
+ items=(
+ ("RIGIFY", "Rigify", "Rigify rig naming and defaults"),
+ ("ARP", "ARP", "Auto-Rig Pro rig naming and collection behavior"),
+ ),
+ default="RIGIFY",
+ )
+
# Character migrator (manual mode)
migrator_mode: BoolProperty(
name="Automatic",
diff --git a/utils/remove_original.py b/utils/remove_original.py
new file mode 100644
index 0000000..47d92f0
--- /dev/null
+++ b/utils/remove_original.py
@@ -0,0 +1,118 @@
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+
+"""Resolve which collection to remove when deleting the original character armature."""
+
+import bpy
+
+
+def _parent_collection(scene, coll):
+ """Return the parent of coll in the scene tree, or None if not found."""
+ if coll is None:
+ return None
+ master = scene.collection
+ if coll in master.children:
+ return master
+ for p in bpy.data.collections:
+ if coll in p.children:
+ return p
+ return None
+
+
+def _depth_from_scene_root(scene, coll):
+ """Depth: number of steps walking up from coll until scene.collection (deeper = larger)."""
+ d = 0
+ cur = coll
+ while cur is not None and cur != scene.collection:
+ d += 1
+ cur = _parent_collection(scene, cur)
+ return d
+
+
+def _walk_up_chain(scene, coll):
+ """Return [inner, ..., top] where top is a direct child of scene.collection (or highest ancestor)."""
+ chain = []
+ cur = coll
+ while cur is not None and cur != scene.collection:
+ chain.append(cur)
+ cur = _parent_collection(scene, cur)
+ return chain
+
+
+def _deepest_users_collection(scene, armature):
+ """Among armature.users_collection, pick the most nested (max depth) as the inner anchor."""
+ colls = list(getattr(armature, "users_collection", []) or [])
+ if not colls:
+ return None
+ best = colls[0]
+ best_d = _depth_from_scene_root(scene, best)
+ for c in colls[1:]:
+ d = _depth_from_scene_root(scene, c)
+ if d > best_d:
+ best_d = d
+ best = c
+ return best
+
+
+def _library_ptr(lib):
+ """Stable id for comparing library datablocks."""
+ return lib.filepath if lib else None
+
+
+def _pick_remove_target_from_chain(orig, chain, rig_family):
+ """
+ From walk-up chain [inner..top], choose one collection to remove.
+ Prefer the topmost under scene (chain[-1]) so nested linked setups remove the whole instance.
+ If orig is linked, prefer the outermost chain entry whose library matches orig.library.
+ """
+ if not chain:
+ return None
+ orig_lib = _library_ptr(getattr(orig, "library", None))
+
+ # Linked armature: outermost in chain from same library file (reversed = top -> inner)
+ if orig_lib:
+ for c in reversed(chain):
+ if _library_ptr(getattr(c, "library", None)) == orig_lib:
+ return c
+
+ # Local override: prefer outermost collection that participates in override hierarchy
+ if getattr(orig, "override_library", None):
+ for c in reversed(chain):
+ if getattr(c, "override_library", None) is not None:
+ return c
+
+ # Rigify: optional name hint — prefer collection whose name matches base armature name
+ if rig_family == "RIGIFY":
+ name = orig.name
+ base = name.replace("_Rigify", "").replace(".001", "").rstrip("0123456789.")
+ for c in reversed(chain):
+ if c.name == base or base in c.name or c.name in name:
+ return c
+
+ # Default: top of chain (direct child of scene.collection subtree root)
+ return chain[-1]
+
+
+def resolve_collection_for_remove_original(orig, rig_family, scene):
+ """
+ Return a collection to remove for Remove Original, or None to fall back to object-only removal.
+
+ Walks up from the deepest users_collection so nested linked rigs remove the outer instance,
+ not an inner linked child collection.
+
+ rig_family: 'RIGIFY' | 'ARP' (ARP skips Rigify name heuristics in _pick_remove_target_from_chain).
+ """
+ if not orig or orig.type != "ARMATURE" or orig.name not in bpy.data.objects:
+ return None
+
+ inner = _deepest_users_collection(scene, orig)
+ if inner is None:
+ return None
+
+ chain = _walk_up_chain(scene, inner)
+ if not chain:
+ return None
+
+ return _pick_remove_target_from_chain(orig, chain, rig_family)