add-gpu-op

Add a new operation to the OpenVINO GPU plugin — OpenCL kernel design, oneDNN-backed paths, sub-group/LWS tuning, and functional tests.

Quality

66%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./.github/agents-prototype/skills/add-gpu-op/SKILL.md

Quality

Content

65%

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

The SKILL.md body is exceptionally concise and well-structured as an overview, but it offloads all executable detail to referenced files that are not actually present in the bundle, and lacks explicit validation checkpoints in its workflow.

Suggestions

Ship the referenced step files (step0-plan.md, step1-hardware-analysis.md, etc.) in ./references/ so the one-level-deep navigation resolves.

Add at least one explicit validation checkpoint, e.g. 'After step 3c, confirm layer tests pass before proceeding to profiling.'

Include a minimal concrete example (a kernel skeleton or registration snippet) directly in the body so the skill is actionable even before opening a referenced file.

Dimension	Reasoning	Score
Conciseness	The body is lean: a short 'When to use' line and a compact step table with no concept explanation or padding — every token earns its place.	3 / 3
Actionability	It sequences steps and names target files, but delegates all concrete executable guidance (kernel code, test commands, profiling) to referenced files, so the body itself describes rather than instructs.	2 / 3
Workflow Clarity	Steps are clearly ordered (0a through 5) with per-step purpose and artifact handoff, but there are no explicit validation/checkpoint steps for the risky kernel-test-profile cycle.	2 / 3
Progressive Disclosure	The body uses the ideal overview-plus-one-level-deep-references pattern with a well-signaled navigation table, but none of the referenced files (step0-plan.md, orchestrator.md, etc.) exist, so navigation to detail is broken.	2 / 3
	Total	9 / 12 Passed

Description

67%

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

The description is specific and well-scoped to a clear niche, but it omits an explicit 'Use when...' trigger clause, so it answers 'what' well while leaving 'when' only implied.

Suggestions

Add an explicit trigger clause, e.g. 'Use when implementing GPU plugin support for a new op spec from the Core OpSpec agent.'

Include one or two common user-level terms (e.g. 'GPU kernel', 'OpenVINO op') alongside the technical jargon to broaden natural-trigger coverage.

Dimension	Reasoning	Score
Specificity	Names multiple concrete actions — 'OpenCL kernel design', 'oneDNN-backed paths', 'sub-group/LWS tuning', and 'functional tests' — rather than vague language.	3 / 3
Completeness	Clearly states what the skill does but lacks any 'Use when...' clause or explicit trigger guidance, which caps completeness at 2 per the rubric.	2 / 3
Trigger Term Quality	Uses relevant domain terms a GPU-plugin contributor would say ('OpenCL', 'oneDNN', 'sub-group/LWS') but they skew toward niche jargon and omit common user-level variations.	2 / 3
Distinctiveness Conflict Risk	The tightly scoped niche ('OpenVINO GPU plugin') with its specialized terms makes it unlikely to trigger for unrelated skills.	3 / 3
	Total	10 / 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
relative_links	Relative link issues: 11 missing	Warning

	Total	15 / 16 Passed

Repository: openvinotoolkit/openvino
Path: .github/agents-prototype/skills/add-gpu-op/SKILL.md
Commit: 4c20980

Reviewed: 4 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.