MAX-786/claude-3d-harness / library/cc/quality-refinement-autoloop
quality-refinement-autoloop
VerifiedRun a refinement loop when a Blender result is subpar, user expectations are not met, validation fails, or repeated issues reveal a missing method. The loop freezes the work, captures evidence, diagnoses the failing dimension, decides which existing skill repairs it, records the lesson, and repairs the artifact from the correct baseline.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 3.4 KB |
| scripts/ralph_autoloop_plan.py | script | 2.6 KB |
scripts/ralph_autoloop_plan.py
View on GitHub#!/usr/bin/env python3
"""Create a generic quality-refinement autoloop plan from user feedback."""
import argparse, json, re
KEYWORDS = {
'geometry': ['shape','silhouette','wireframe','landmark','proportion','does not fit','model behind'],
'multiview_depth': ['side','back','top','depth','thickness','z-axis','edge-on'],
'uv_texture': ['texture','uv','atlas','decal','stretched','not fill','not aligned'],
'surface_coverage': ['side texture','back texture','whole model','surface','sidewall','closed surface'],
'look_lighting': ['lighting','glow','material','color','too dark','too bright','look'],
'animation': ['animation','motion','morph','spin','loop','speed','frame'],
'export_truth': ['glb','export','runtime','mp4','sequence','viewer'],
'orchestration': ['same issue','repeated','loop','skill','harmonization','handoff'],
}
SKILL_HINTS = {
'geometry': ['texture-driven-mesh-fitting','landmark-fit-repair','contour-to-mesh'],
'multiview_depth': ['orthographic-registration','multiview-constraint-solver','multiview-fit-loop'],
'uv_texture': ['blender-uv-texturing','atlas-uv-fitting'],
'surface_coverage': ['closed-surface-uv-coverage','blender-uv-texturing'],
'look_lighting': ['reference-look-calibration','blender-materials','blender-lighting'],
'animation': ['texture-state-animation','orbital-hud-motion','animation-quality-gate'],
'export_truth': ['blender-export','animation-quality-gate'],
'orchestration': ['blender-skill-harmonizer','quality-refinement-autoloop'],
}
def main():
ap=argparse.ArgumentParser()
ap.add_argument('--feedback', required=True)
ap.add_argument('--artifact', default='')
ap.add_argument('--out')
args=ap.parse_args()
text=(args.feedback+' '+args.artifact).lower()
scores={k:sum(1 for kw in kws if kw in text) for k,kws in KEYWORDS.items()}
ranked=[k for k,v in sorted(scores.items(), key=lambda kv: kv[1], reverse=True) if v]
if not ranked: ranked=['orchestration']
plan={
'schema':'quality_refinement_autoloop_plan.v1',
'feedback':args.feedback,
'artifact':args.artifact,
'primary_failure':ranked[0],
'failure_dimensions':ranked,
'candidate_skills':sorted({s for d in ranked for s in SKILL_HINTS[d]}),
'phases':['freeze_baseline','capture_evidence','diagnose','decide_skill_gap','record_lesson','repair_product'],
'hard_gates':['do_not_rebuild_before_skill_gap_decision','do_not_edit_skill_files','run_validation_report_before_final']
}
txt=json.dumps(plan,indent=2)
if args.out: open(args.out,'w').write(txt)
print(txt)
if __name__=='__main__': main()
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: 471k | 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 |