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

78%

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

Fix and improve this skill with Tessl

tessl review fix ./packages/python/skills/ax-python-refine/SKILL.md

Quality

Content

72%

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

A compact, well-structured API-reference skill that is token-efficient and well-organized with sensible progressive disclosure. It is held back by a core code sketch that relies on undefined placeholders and an API surface list lacking signatures, plus the absence of a clear sequenced workflow for its multiple use cases.

Suggestions

Make the Core Pattern example copy-paste ready by defining or labeling the placeholder variables (reflection_client, request, evaluator), e.g. showing how a reflection client and evaluator are constructed.

Expand 'Relevant API Surface' with minimal signatures or one-line usage for each optimizer (AxGEPA, AxPlaybook, AxBootstrapFewShot) so Claude can pick the right one without loading external docs.

Add a short sequenced decision flow under 'When To Use' mapping each scenario to the optimizer/pattern to use, giving the multi-step intent an explicit order.

Dimension	Reasoning	Score
Conciseness	The body is lean — short bullets and a minimal code sketch — and assumes Claude's competence without explaining basic Python or library concepts. Not a 2 because there is no padded or unnecessary explanation; every line carries information.	3 / 3
Actionability	Provides a concrete import and call shape ('from axllm import AxGEPA', 'engine.optimize(request, evaluator)') and lists real API names, but the core sketch uses undefined placeholders ('reflection_client', 'request', 'evaluator') and 'Relevant API Surface' gives names without signatures. Not a 3 because the example is not copy-paste ready; not a 1 because real, specific API surface is named.	2 / 3
Workflow Clarity	'When To Use' lists scenarios and 'Core Pattern' gives a canonical call, but there is no sequenced multi-step process or decision flow for choosing among optimizers. Not a 3 because, although the skill is compact, the multi-scenario guidance is not unambiguously sequenced; not a 1 because a clear core action is present.	2 / 3
Progressive Disclosure	A well-organized, sub-50-line overview with clear sections that points to one-level-deep detail materials ('API.md', 'axir-api.json', 'axir-capabilities.json', 'examples/'). Not a 2 because content is appropriately split into a concise overview rather than inlined as a monolithic wall; no bundle files were present to verify reference resolution.	3 / 3
	Total	10 / 12 Passed

Description

85%

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, well-scoped description that clearly states both purpose and an explicit 'Use when' trigger with several concrete capabilities. Its only weakness is moderately technical phrasing that misses some natural user terms a layperson might say.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'reward-scored generation', 'iterative candidate improvement', 'evaluator feedback', 'optimizer-backed refinement patterns' — rather than vague language. Not a 2 because it enumerates several specific operations, not just a domain plus one action.	3 / 3
Completeness	Explicitly states both what ('reward-scored generation, iterative candidate improvement, evaluator feedback, and optimizer-backed refinement patterns') and when ('Use when writing Python code with `axllm`') with an explicit 'Use when' trigger clause. Not a 2 because the 'when' is explicit, not merely implied.	3 / 3
Trigger Term Quality	Includes relevant keywords ('Python code', 'axllm', 'evaluator feedback') but leans technical with phrases like 'reward-scored generation' and 'optimizer-backed refinement', missing common layperson variations such as 'improve prompts' or 'optimize LLM outputs'. Not a 3 because coverage of natural user terms is incomplete.	2 / 3
Distinctiveness Conflict Risk	Scoped to a specific package ('writing Python code with `axllm`') with distinct refinement/optimization triggers, making it unlikely to fire for unrelated skills. Not a 2 because the package name pins a clear niche rather than overlapping generically.	3 / 3
	Total	11 / 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
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: fe8baf8

Reviewed: 5 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.