webnn

Implements and debugs browser Web Neural Network API integrations in JavaScript or TypeScript web apps. Use when adding navigator.ml checks, MLContext creation, MLGraphBuilder flows, device selection, tensor dispatch and readback, or explicit fallback paths to ONNX Runtime Web or other local runtimes. Don't use for model training, server-side ML inference, or cloud AI APIs.

Quality

96%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

—

The risk profile of this skill

Quality

Content

92%Weight 40%Scale 1-3

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

A well-structured, lean procedural skill with strong sequencing, validation checkpoints, and concrete API guidance. Its one real weakness is progressive disclosure: the body points to six reference/template/script files that are not present in the bundle, breaking navigation.

Suggestions

Ship the missing bundle files (references/webnn-reference.md, examples.md, compatibility.md, troubleshooting.md, assets/webnn-runtime.template.ts, scripts/find-webnn-targets.mjs) so the signaled references actually resolve.

If any referenced file is not intended to exist, remove the pointer or inline the essential content so Claude is not sent to a dead path.

Add a short bundle manifest or one-line note near the first reference so it is obvious which files are loadable and in what order to consult them.

Dimension	Reasoning	Score
Conciseness	The body is lean and instruction-dense with no explaining of concepts Claude already knows (no 'WebNN is a…' preamble); every line is an action or a reference pointer, so it earns the 'lean and efficient; every token earns its place' anchor rather than the 'mostly efficient but could be tightened' 2.	3 / 3
Actionability	Provides real executable commands ('node scripts/find-webnn-targets.mjs .') and concrete API calls ('context.writeTensor()', 'context.dispatch()', 'await context.readTensor()', 'powerPreference: "high-performance"') rather than pseudocode, matching the 'fully executable code/commands' anchor.	3 / 3
Workflow Clarity	Five clearly sequenced steps with an explicit validation phase (Step 5: smoke test, 'Run the workspace build, typecheck, or tests after editing'), stop/ask guard conditions in Step 1, and feedback loops in Error Handling — satisfying the 'clear sequence with explicit validation steps' anchor.	3 / 3
Progressive Disclosure	References are clearly signaled and one level deep ('Read references/webnn-reference.md', 'Read assets/webnn-runtime.template.ts'), but the bundle directories references/, scripts/, and assets/ do not exist, so every reference is a dead link and navigation fails; this caps the score below 3 even though organization is better than the 'monolithic wall / poor organization' anchor 1.	2 / 3
	Total	11 / 12 Passed

Description

100%Weight 40%Scale 1-3

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, third-person description that states concrete capabilities, gives explicit 'Use when…' triggers with natural terms, and disambiguates itself with clear 'Don't use for…' negatives. It satisfies all four dimensions at the top anchor.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Implements and debugs browser Web Neural Network API integrations', 'navigator.ml checks, MLContext creation, MLGraphBuilder flows, device selection, tensor dispatch and readback…fallback paths to ONNX Runtime Web' — matching the 'lists multiple specific concrete actions' anchor; not a 2 because it goes well beyond naming a domain plus a few actions.	3 / 3
Completeness	Explicitly answers 'what' ('Implements and debugs browser Web Neural Network API integrations in JavaScript or TypeScript web apps') and 'when' ('Use when adding navigator.ml checks, MLContext creation…'), plus explicit 'Don't use for…' negatives; not a 2 because the trigger guidance is explicit rather than implied.	3 / 3
Trigger Term Quality	Uses natural terms a user would say — 'Web Neural Network', 'navigator.ml', 'MLContext', 'MLGraphBuilder', 'tensor dispatch and readback', 'ONNX Runtime Web', 'fallback' — giving good coverage of common variations rather than only jargon or a single keyword.	3 / 3
Distinctiveness Conflict Risk	Occupies a clear niche (browser WebNN API) with distinct triggers, and the 'Don't use for model training, server-side ML inference, or cloud AI APIs' clause separates it from general and cloud ML skills, making wrong-skill activation unlikely.	3 / 3
	Total	12 / 12 Passed

Validation

93%

Warnings & errors only

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

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
referenced_paths_exist	Referenced path issues: 11 missing	Warning

	Total	15 / 16 Passed

Repository: webmaxru/web-ai-agent-skills
Path: skills/webnn/SKILL.md
Commit: f4c5a23

Reviewed: 13 days ago

Table of Contents

Discovery Implementation Validation

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.