CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

iwsdk-compose-scene

Compose and review IWSDK scenes from existing manifest assets using scene JSON, placement, scale, contacts, prefabs, patterns, lighting, environment, cameras, and final framing. Does not author geometry or materials; create or repair an individual 3D asset with iwsdk-build-model.

Quality

85%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

88%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.

A dense, highly actionable skill body: complete executable templates, exact commands, bounded budgets, and a gated workflow with explicit finish conditions, with almost no generic filler. Its only weaknesses are mild repetition and a two-level reference chain through extended-workflow.md.

DimensionReasoningScore

Conciseness

Nearly every sentence states a non-obvious project contract ("Scene JSON has one renderable content kind: `asset`", "Directional lights emit along local `-Z`", "there is no `color` field") with no general-knowledge padding, but there is minor repetition — the extended-workflow trigger list ("imports, prefabs, patterns, ... schema errors, image reconstruction") appears twice — and a few long prose runs (the editor/runtime paragraph, the screenshot-readiness paragraph) that could be tightened. Anchor 4 (efficient, minor trimming possible) fits better than 5 because of that duplicated and slightly padded material.

4 / 5

Actionability

Fully executable throughout: a complete copy-paste scene JSON template ("Use this complete minimal shape instead of looking up the common schema"), exact CLI invocations with flags (`npx @iwsdk/cli scene render-file --input-json ... --output-file ...`), the named MCP tool surface, concrete component payloads (`HemisphereLight: {skyColor, groundColor, intensity}`), and numeric budgets (512×512 first render, 384×384 verification, three renders, two image reads). This matches the anchor for copy-paste ready guidance covering the common cases.

5 / 5

Workflow Clarity

The Fast Path is a numbered 7-step sequence with explicit validation checkpoints and a feedback loop (render hero → "make at most one coherent visual correction" → "verify that correction at 384×384"), explicit finish gates ("open the file, confirm clean state, and run the application build", "stop"), and failure-handling guidance for blocked visual verification. The Extended Workflow is delegated to a reference whose own structure (Specify → Inventory → Compose → Validate → Review → Finish) was verified. Not 4: checkpoints, error recovery, and stop conditions are all explicit rather than implicit.

5 / 5

Progressive Disclosure

Good structure: the common path is deliberately inline with instructions not to load references for it ("For this common path, do not read the references"), and both body links ([extended-workflow.md](references/extended-workflow.md), [environment-recipes.md](references/environment-recipes.md)) resolve to real files. Not 5: references are not one level deep — SKILL.md routes to extended-workflow.md, which itself links five further references (scene-format.md, text-intake.md, image-intake.md, scene-composition.md, scene-review.md), a two-level chain; it is clearly signaled and intentional routing, but it exceeds the one-level-deep ideal.

4 / 5

Total

18

/

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 specific, well-bounded description with strong distinctiveness and concrete capability enumeration, undermined by the absence of any explicit trigger guidance telling Claude when to invoke it. Adding a 'Use when...' clause would raise completeness and trigger-term coverage together.

Suggestions

Add an explicit trigger clause, e.g. "Use when composing, placing, lighting, or framing IWSDK scene JSON from existing manifest assets, or when reviewing a composed scene."

Include user-facing synonyms or file extensions in the trigger terms, such as ".iwsdk.scene.json" or "scene files", so requests phrased around the artifact also match.

DimensionReasoningScore

Specificity

"Compose and review IWSDK scenes" plus an enumerated list of concrete operations ("placement, scale, contacts, prefabs, patterns, lighting, environment, cameras, and final framing") covers the domain comprehensively, matching the anchor for multiple specific concrete actions. Not 4: no meaningful coverage gaps within the stated niche — even the exclusion ("Does not author geometry or materials") is explicit.

5 / 5

Completeness

The "what" is clear ("Compose and review IWSDK scenes from existing manifest assets") but there is no "Use when..." clause or equivalent explicit trigger guidance; the routing sentence "create or repair an individual 3D asset with iwsdk-build-model" is a boundary, not a trigger. Per the rubric guideline, a missing explicit 'when' caps completeness at 3, so it cannot score 4.

3 / 5

Trigger Term Quality

Natural terms a user would say are present — "compose", "scene", "placement", "lighting", "cameras", "prefabs", "review" — giving good keyword coverage. Not 5: no synonyms or file extensions (e.g. ".iwsdk.scene.json" / "scene file") to broaden matching, and "manifest assets" is internal jargon rather than user vocabulary.

4 / 5

Distinctiveness Conflict Risk

"IWSDK scenes from existing manifest assets using scene JSON" carves out a clear niche, and the explicit contrast with "iwsdk-build-model" (geometry/materials vs. composition) minimizes conflict with the sibling skill. Not 4: the boundary is stated in the description itself, not merely inferable.

5 / 5

Total

17

/

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

facebook/immersive-web-sdk

Path

packages/core/agent-guidance/claude/.claude/skills/iwsdk-compose-scene/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.