liger-autopatch

Adds Liger Kernel support for a new HuggingFace Transformers model, or modifies existing monkey-patching. Generates lce_forward, monkey-patch function, tests, and README entry. Use when adding a new model to Liger Kernel, when a user asks to patch an unsupported model, when extending MODEL_TYPE_TO_APPLY_LIGER_FN, or when modifying/updating/fixing an existing monkey-patch (e.g., adding a new kernel to an already-supported model, fixing instance patching, updating a patch for upstream HF changes).

Quality

92%

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

Quality

Content

85%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 well-structured, staged pipeline skill with strong workflow clarity and clean progressive disclosure through one-level-deep references. Minor conciseness and actionability gaps come from redundant inline test listings and delegation to external workflow files.

Suggestions

Remove the fully enumerated test command list in Modify Mode Stage 3 and rely on validator.md to avoid duplication with Create Mode's Validate stage.

Consider inlining a one-line summary of what each delegated workflow (model-analyzer, code-generator) produces so the body remains actionable even if a referenced file is unavailable.

Add a brief note clarifying where the referenced files (model-analyzer.md, templates/, etc.) live relative to SKILL.md so navigation is unambiguous.

Dimension	Reasoning	Score
Conciseness	Lean and well-organized with no padding or explanation of known concepts, but Modify Mode Stage 3 redundantly enumerates all six convergence test commands that are already covered by the referenced validator.md.	4 / 5
Actionability	Provides concrete executable commands (pytest invocations with -k filters, make checkstyle) and enumerates 13 specific file paths, but the Analyze and Generate stages delegate to external workflow files rather than giving inline steps.	4 / 5
Workflow Clarity	A clearly sequenced three-stage pipeline with explicit human checkpoints between every stage, a mandatory Validate stage with minimum test list, and a feedback loop ('Retries up to 3 times on failure').	5 / 5
Progressive Disclosure	The body is a concise overview pointing one level deep to clearly signaled references — model-analyzer.md, decision-matrix.md, code-generator.md, validator.md, examples/, and templates/ — with no nested reference chains.	5 / 5
	Total	18 / 20 Passed

Description

100%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, specific description that clearly states concrete actions and provides explicit, varied 'Use when...' trigger phrases for both create and modify scenarios. It uses third person throughout and avoids vague language.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Generates lce_forward, monkey-patch function, tests, and README entry' — covering both adding support for a new model and modifying existing monkey-patching.	5 / 5
Completeness	Explicitly answers both what ('Adds Liger Kernel support... Generates lce_forward, monkey-patch function, tests, and README entry') and when ('Use when adding a new model... when a user asks to patch an unsupported model...').	5 / 5
Trigger Term Quality	Comprehensive natural trigger phrases with synonyms — 'adding a new model', 'patch an unsupported model', 'modifying/updating/fixing', 'adding a new kernel to an already-supported model' — alongside the technical symbol MODEL_TYPE_TO_APPLY_LIGER_FN.	5 / 5
Distinctiveness Conflict Risk	Occupies a clear niche — Liger Kernel monkey-patching for HuggingFace Transformers models — with distinct triggers and minimal overlap risk with other skills.	5 / 5
	Total	20 / 20 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: 10 missing	Warning

	Total	15 / 16 Passed

Repository: linkedin/Liger-Kernel
Path: .agents/skills/liger-autopatch/SKILL.md
Commit: 7777995

Reviewed: 2 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.