ax-python-refine

Use when writing Python code with `axllm` for reward-scored generation, iterative candidate improvement, evaluator feedback, and optimizer-backed refinement patterns.

Quality

82%

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

78%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 lean, information-dense body that adds only what Claude could not know about the generated axllm package. The Core Pattern is close to executable and guardrails are decision-useful, but referenced materials are not shipped in the bundle and no explicit ordered workflow or verification checkpoint is given.

Suggestions

Ship or clearly locate the referenced materials (API.md, axir-api.json, axir-capabilities.json, examples/) inside the skill bundle — e.g., a references/ directory — or state explicitly that they live in the generated package repo, so every reference resolves.

Make the Core Pattern example runnable: define or annotate 'reflection_client', 'request', and 'evaluator', or point to the specific examples/ file to copy first.

Add a short ordered workflow with a validation checkpoint, e.g.: 1. check axir-capabilities.json for the available optimizer surface, 2. start from the closest examples/ case, 3. verify call signatures against axir-api.json before writing new code.

Dimension	Reasoning	Score
Conciseness	Telegraphic bullets and one small code block with zero padding; every line carries package-specific facts Claude cannot know (runtime profiles, manifests, transport modes). Assumes Claude's competence throughout.	5 / 5
Actionability	The Core Pattern is real API-shaped code but uses undefined placeholders ('reflection_client', 'request', 'evaluator'), so it is not copy-paste ready; pointers to 'API.md', 'axir-api.json', and 'examples/' fill most gaps.	4 / 5
Workflow Clarity	Clear single-purpose guidance with decision rules in Guardrails (examples first, provider-api vs no-key transport, AxIR as source of truth), but no explicit sequence or validation checkpoint such as verifying signatures against axir-api.json before writing.	4 / 5
Progressive Disclosure	The body is well-sectioned, but it references 'API.md', 'axir-api.json', 'axir-capabilities.json', and 'examples/' that are not present in the skill bundle, and these references are listed as package facts rather than clearly signaled navigation links.	3 / 5
	Total	16 / 20 Passed

Description

78%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 description with an explicit 'Use when' clause, multiple concrete capability phrases, and a distinct niche anchored on the axllm package name. Its main weakness is that the 'when' clause mirrors the 'what' instead of adding separate trigger scenarios, and a few natural synonym triggers are absent.

Dimension	Reasoning	Score
Specificity	The description lists several concrete capability areas — 'reward-scored generation, iterative candidate improvement, evaluator feedback' — though 'optimizer-backed refinement patterns' is abstract. Minor coverage gaps keep it below the comprehensive anchor.	4 / 5
Completeness	Both 'what' and 'when' are present via the explicit 'Use when writing Python code with axllm' clause. The 'when' largely restates the 'what' rather than adding distinct trigger situations, so it is not a 5.	4 / 5
Trigger Term Quality	Good natural keyword coverage: 'writing Python code', 'iterative candidate improvement', 'evaluator feedback', 'refinement'. A few natural terms users might say are missing (e.g., 'optimize prompts', 'GEPA', 'refine outputs').	4 / 5
Distinctiveness Conflict Risk	Anchored to the niche package name 'axllm' and optimizer-specific terminology ('reward-scored generation', 'optimizer-backed refinement'), giving a clear niche with distinct triggers and minimal conflict risk.	5 / 5
	Total	17 / 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: packages/python/skills/ax-python-refine/SKILL.md
Commit: 3f630a0

Reviewed: 3 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.