MAX-786/claude-3d-harness / library/cc/fit-repair-optimizer
fit-repair-optimizer
VerifiedTurn multiview fit reports into a sequential or parallel repair queue for aligning Blender products to source-of-truth templates. Use after validation shows disalignment, when the agent must iteratively fix wireframe, texture, lighting, or projection mismatches, and when a recurring failure should stop the rebuild loop for a diagnosis.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 3.2 KB |
| scripts/fit_repair_queue.py | script | 5.2 KB |
SKILL.md
View on GitHub---
name: fit-repair-optimizer
description: Turn multiview fit reports into a sequential or parallel repair queue for aligning Blender products to source-of-truth templates. Use after validation shows disalignment, when the agent must iteratively fix wireframe, texture, lighting, or projection mismatches, and when a recurring failure should stop the rebuild loop for a diagnosis.
when_to_use: Iterative source-of-truth alignment repair, choosing sequential vs parallel correction order, generating fit repair queues, stopping on contradictory templates, or coordinating geometry/UV/lighting fixes from validation reports.
allowed-tools: Read Bash Glob Grep mcp__blender__execute_blender_code mcp__blender__get_scene_info mcp__blender__get_object_info
---
# Fit Repair Optimizer
This skill converts validation failures into an executable repair queue. The source templates are the contract; the current model is disposable.
## Decision: sequential or parallel?
Use **sequential repair** when failures are coupled:
1. Canonical view / source conflict unresolved.
2. Front silhouette or part count fails.
3. Geometry dimensions fail before UVs can be trusted.
4. Texture region boundaries depend on geometry changes.
5. Lighting/look depends on final material placement.
Use **parallel repair** only for independent tracks with disjoint write scopes:
- geometry recipe/placement tuning;
- atlas-region classification;
- material/look calibration;
- validation tooling improvements.
Do not parallelize two tasks that both edit the same Blender recipe, object transforms, or UV assignment file.
## Mandatory repair order
```
0 source-conflict / multiview-rigidity gate
1 structural part count
2 front silhouette and landmarks
3 side/back/top depth and projection
4 UV/texture region fit
5 material/light/look calibration
6 export and final validation
```
If an earlier stage fails, later stages may be analyzed but must not be finalized.
## Source-conflict gate
Before repair, use `multiview-constraint-solver` to check whether templates are mutually satisfiable. If the same physical axis receives incompatible ratios across views, write a conflict report and require a canonical policy:
- front+side canonical;
- front+top canonical;
- corrected template sheet;
- view-specific cheat renders only, not a rigid 3D model.
## Repair queue schema
Each repair item must include:
- `id`
- `stage`
- `view_scope`
- `failure`
- `measured_delta`
- `proposed_action`
- `write_scope`
- `parallel_group`
- `blocked_by`
- `acceptance_gate`
## Skill-gap rule
If the same failure recurs twice, stop and invoke `quality-refinement-autoloop` before another rebuild. The autoloop must capture evidence, diagnose the missing method and record the lesson, then return here with a new repair queue. It does not edit skill files. Common routing:
- geometry mismatch → `contour-to-mesh`, `orthographic-registration`, or this skill;
- UV/texture mismatch → `atlas-uv-fitting`;
- lighting mismatch → `reference-look-calibration`, then `blender-lighting`;
- validation mismatch → `multiview-fit-loop` / `reference-analysis-validator`.
## Output
Write a project repair plan such as `ALIGNMENT_REPAIR_QUEUE.json` and a human-readable `ALIGNMENT_REPAIR_METHOD.md`.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |