ax-refine

Use this skill when writing or reviewing Ax bestOfN/refine code, reward functions, thresholds, native sample selection, serial attempts, generated advice, and attempt diagnostics.

Quality

93%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

93%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 concise, actionable reference with executable code and clear decision guidance, well-organized into focused sections. Workflow clarity could benefit from an explicit validation/error-recovery sequence.

Suggestions

Add a short validate->fix->retry note for the assertion/reward feedback loops (e.g., how failed addAssert output cycles back into the next attempt) to strengthen workflow clarity.

Clarify the error-recovery path when refine advice fails to apply (adviceApplied: false) so users know the expected next step.

Dimension	Reasoning	Score
Conciseness	The body is lean with no padding or explanations of concepts Claude already knows; every line carries API or decision detail and assumes Claude's competence.	5 / 5
Actionability	Provides copy-paste ready TypeScript for bestOfN, refine, and a reward function, plus concrete API rules covering the common cases.	5 / 5
Workflow Clarity	Decision guidance (bestOfN vs refine, when to use serial) and assertion boundaries are clearly sequenced, but there is no explicit error-recovery or validation checkpoint loop.	4 / 5
Progressive Disclosure	Well-organized sections with no nesting; no bundle files exist so the self-contained under-50-line body fits the simple-skill exception for well-structured content.	5 / 5
	Total	19 / 20 Passed

Description

92%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 with an explicit trigger clause and concrete capability list that clearly distinguishes its niche. Trigger term coverage is good but slightly jargon-leaning.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete capability areas ('bestOfN/refine code, reward functions, thresholds, native sample selection, serial attempts, generated advice, and attempt diagnostics'), giving comprehensive coverage of the skill's scope.	5 / 5
Completeness	Explicit 'Use this skill when...' trigger clause plus a concrete 'what' list clearly and explicitly answer both what the skill does and when to use it.	5 / 5
Trigger Term Quality	Includes natural developer phrases ('writing or reviewing', 'reward functions', 'thresholds', 'serial attempts', 'generated advice') with good coverage, but leans technical and lacks broader synonyms or common variations.	4 / 5
Distinctiveness Conflict Risk	The Ax bestOfN/refine niche is highly specific with distinct triggers, making it unlikely to fire for unrelated skills.	5 / 5
	Total	19 / 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
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: ax-llm/ax
Path: website/static/typescript/.well-known/agent-skills/ax-refine/SKILL.md
Commit: 1abea97

Reviewed: ~18 hours 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.