CtrlK
BlogDocsLog inGet started
Tessl Logo

add-core-op

Adds a core operator to the OpenVINO toolkit. Use when asked to implement a new operation into OpenVINO.

60

Quality

76%

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

Fix and improve this skill with Tessl

tessl review fix ./.github/agents-prototype/skills/add-core-op/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

75%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 solid, dense instruction document: concrete paths, macros, and test locations make it highly actionable, and it wastes almost nothing on concepts Claude already knows. The main improvements are structural — normalizing the heading hierarchy, deduplicating the file-list and class-pattern sections, and making the test step an explicit validate-and-iterate checkpoint.

Suggestions

Normalize the heading hierarchy (currently #, ##, ####, ##### mix, with a second H1 '# Adding a New Operator to OpenVINO' mid-document) so sections are consistently navigable.

Make the test step an explicit feedback loop, e.g., 'Run the type_prop and opset tests after implementation; fix and re-run until green before writing the .rst spec.'

Merge the duplicated file-path/class-pattern material ('Create Header and Source Files' vs 'Files to create/update' and 'Core Op Class Pattern') into one section to save tokens and avoid drift.

DimensionReasoningScore

Conciseness

The body is directive and assumes OpenVINO competence — it never explains what OpenVINO is or how ops work in general, and nearly every line is a rule, path, or code pointer. Minor trimming is possible (the 'Files to create/update' section repeats the .hpp/.cpp paths already given earlier, and duplicated H1 headings '# Skill Instructions' / '# Adding a New Operator to OpenVINO' add padding), which places it at 4 rather than the fully lean anchor 5.

4 / 5

Actionability

Guidance is concrete: exact file paths with globs, code snippets (OPENVINO_OP macro, visitor.on_attribute, OV_OP_SCOPE), specific method names, and precise test directories. It falls short of 5 because the snippets are illustrative skeletons rather than copy-paste-ready code (e.g., the empty class body 'class OPENVINO_API OpName : public Op {}' and the constructor implementation is described but not shown), leaving minor gaps.

4 / 5

Workflow Clarity

A clear 4-step sequence (analysis → update files → tests → specification) is given up front and expanded in ordered sections, and the 'Tests' section enumerates exactly where each test kind goes — a validation checkpoint of sorts. It stays at 4 rather than 5 because there is no explicit run-tests/fix/re-run feedback loop, and the analysis step is named but has no checklist of what to verify.

4 / 5

Progressive Disclosure

No bundle files exist, and the single-file body is a reasonably sized, well-sectioned overview for a moderately complex procedure — so content placement is appropriate for a one-level structure. Minor organization gaps (inconsistent heading levels mixing #, ##, ####, #####; a second H1 mid-document; 'Core Op Class Pattern' partially duplicating earlier sections) keep it at 4 rather than the fully clean-navigation anchor 5.

4 / 5

Total

16

/

20

Passed

Description

66%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 clean, concise, third-person description that clearly states what the skill does and when to use it, with strong distinctiveness. Its main weakness is thin trigger coverage and a one-action statement of what the skill actually involves.

Suggestions

Add 1-2 concrete actions to the 'what' clause, e.g., 'Adds a core operator to the OpenVINO toolkit: defines the op class, registers it in the latest opset, and adds shape inference, tests, and an .rst specification.'

Broaden trigger terms to include natural variations users might say: 'new op', 'custom operator', 'operator support', 'missing operation', 'opset'.

Consider adding a second trigger condition, e.g., 'Use when a model conversion fails because an operation is missing or decomposition is not performant.'

DimensionReasoningScore

Specificity

The description names the domain ('core operator to the OpenVINO toolkit') and one concrete action ('Adds a core operator'), which matches the 'names domain and 1-2 concrete actions' anchor. It does not list several specific actions (e.g., class definition, registration, shape inference, tests, spec), so score 4's 'several specific actions' is not reached.

3 / 5

Completeness

It explicitly answers both: what ('Adds a core operator to the OpenVINO toolkit') and when ('Use when asked to implement a new operation into OpenVINO'). The single trigger clause is narrower and less rich in concrete trigger phrases than the score-5 example, which lists multiple trigger conditions, so it sits at 4.

4 / 5

Trigger Term Quality

Relevant keywords are present ('core operator', 'OpenVINO', 'implement a new operation'), but common variations users would naturally say are missing — e.g., 'new op', 'custom operator', 'opset', 'operator support'. This matches the 'some relevant keywords but missing common variations' anchor rather than 4's 'good keyword coverage with only a few terms missing'.

3 / 5

Distinctiveness Conflict Risk

'OpenVINO' and 'core operator' form a clear niche with distinct triggers; virtually no other skill would match a request to implement a new OpenVINO operation. Minimal conflict risk, matching the top anchor.

5 / 5

Total

15

/

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
openvinotoolkit/openvino
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.