CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

character-animation-3d

Produz personagens e animacoes 3D para uso em runtime. Trigger em: "AccuRIG", "rig de personagem 3D", "retargeting de animacao", "script bpy que roda headless com blender --background --python", "deform bones vs pose bones", "root motion que ta duplicando o deslocamento", "armature.data.bones e armature.pose.bones", "semantic bone mapping pra desacoplar do nome CC_Base_L_Upperarm", "text-to-motion", "MDM ou MoMask pra gerar animacao", "bake de constraints e IK antes de exportar FBX pro Unity", "Blender instalado", "Maya instalado", "ZBrush instalado", "sculpt", "retopologia", "XGen groom", "mayapy", "ZScript", "GoZ", "polypaint".

Quality

87%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

92%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

Excellent content: highly actionable with copy-paste commands, exact math, and verified complete scripts in references; a clearly sequenced pipeline with explicit validation checkpoints and a fix-rebake feedback loop; and exemplary progressive disclosure with a navigation table mapping topics to real, one-level-deep reference files. The only notable inefficiency is some duplication between prose sections and the Hard Rules table, plus a license table that could live in a reference file.

DimensionReasoningScore

Conciseness

Dense, expert-level content with no explanations of concepts Claude already knows — every section carries non-obvious domain knowledge (data vs pose bones, delta-relative retargeting math, root-motion ownership). Not 5 because of deliberate duplication between prose sections and the 'Regras Duras' table (root motion rule, Deep Search, raw location copy each stated twice) and the 14-row license table in 'Fontes' that reads like reference material inlined in the overview. Not 3 because the prose is tight and the duplication is a defensible summary pattern for the costliest errors.

4 / 5

Actionability

Fully concrete: copy-paste CLI invocation ("blender --background --python pipeline.py -- --character x.fbx --motion y.bvh --output z.fbx"), exact bake call ("bpy.ops.nla.bake(visual_keying=True, clear_constraints=True, bake_types={\"POSE\"})"), exact formulas for retarget delta and root-motion scaling, and the explicit essential-bones list. Verified that the referenced files contain the complete implementations (BONE_ALIASES, find_armature, retarget_frame), covering the common pipeline cases end to end.

5 / 5

Workflow Clarity

The multi-stage pipeline (import -> mapping -> retarget -> bake -> export) is clearly sequenced via the layered-architecture diagram and per-stage sections, with explicit validation throughout: the mapping fails loudly (RuntimeError on missing essential bones), a post-retarget checkpoint with a fix-and-rebake feedback loop ("Corrigir contact points/proporcao e re-bakear antes de exportar"), and a completion-evidence checklist covering every stage. This matches the anchor-5 pattern of explicit validation steps plus error-recovery loops.

5 / 5

Progressive Disclosure

Textbook structure: a 220-line overview with an explicit on-demand loading instruction ("carregar so o arquivo do topico relevante a task") and a subject-to-file navigation table; all seven cited reference files exist, contain the claimed content (verified), and are one level deep with no nested references. Not lower despite the minor stale count ("nao os quatro de cara" — there are now seven files), which does not hinder navigation.

5 / 5

Total

19

/

20

Passed

Description

75%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description with an explicit trigger list containing unusually natural user phrasings (bug descriptions, API names, CLI invocations) and a clearly distinct niche. The main weakness is the thin leading capability statement, which says only 'produces 3D characters and animations for runtime' without naming the pipeline stages the skill actually covers, and a few missing natural terms (file extensions, major tool names like Mixamo/Rokoko/DeepMotion).

Suggestions

Expand the opening 'what' sentence to name the pipeline stages (e.g., 'Certifica rigs com AccuRIG, seleciona/gera motion com IA, retarget e bake no Blender headless, exporta FBX/glTF/Alembic para runtime') so the capability statement is comprehensive rather than a single broad clause.

Add missing natural trigger terms: file extensions (.fbx, .bvh, .glb, .abc) and tool names users commonly mention (Mixamo, Rokoko, DeepMotion, Auto-Rig Pro).

Qualify broad DCC triggers like "sculpt", "retopologia", and "Blender/Maya/ZBrush instalado" so they bind to character-pipeline contexts (e.g., "sculpt de personagem no ZBrush") and reduce overlap risk with general modeling/DCC skills.

DimensionReasoningScore

Specificity

Concrete actions are named via trigger phrases such as "retargeting de animacao", "script bpy que roda headless com blender --background --python", "semantic bone mapping", and "bake de constraints e IK antes de exportar FBX", but the leading capability statement "Produz personagens e animacoes 3D para uso em runtime" is broad and does not by itself comprehensively describe the pipeline. Not 5 because the explicit 'what' sentence lacks the breadth of the anchor-5 example; not 3 because multiple concrete actions beyond 1-2 are explicitly stated.

4 / 5

Completeness

Both parts are present and explicit: what ("Produz personagens e animacoes 3D para uso em runtime") and an explicit when-clause ("Trigger em: ...") with concrete trigger phrases, so it exceeds the anchor-4 example where 'when' is generic. Not 5 because the 'what' is a single broad sentence that omits core capabilities (rig certification, AI motion selection, retarget/bake) that the triggers only imply.

4 / 5

Trigger Term Quality

Excellent natural coverage including bug phrasings users would actually type ("root motion que ta duplicando o deslocamento"), API paths ("armature.data.bones e armature.pose.bones"), and tool-specific terms (AccuRIG, MDM, MoMask, mayapy, ZScript, GoZ, XGen groom). Not 5 because common variations are missing: file extensions (.fbx, .bvh, .glb) and major tool names covered by the skill (Mixamo, Rokoko, DeepMotion, Auto-Rig Pro).

4 / 5

Distinctiveness Conflict Risk

Clear niche (3D character animation pipeline) with highly distinctive triggers (AccuRIG, CC_Base_L_Upperarm, GoZ, mayapy, "deform bones vs pose bones") that are unlikely to fire for the wrong skill. Not 5 because several triggers are broad DCC terms ("sculpt", "retopologia", "Blender instalado", "Maya instalado", "polypaint") that could overlap with a general modeling/DCC skill.

4 / 5

Total

16

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

felvieira/claude-skills-fv

Path

skills/68-character-animation-3d/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.