ax-python-gepa

Use when writing Python code with `axllm` for GEPA, Pareto tradeoffs, reflection clients, metric budgets, optimizer state, and artifacts.

Quality

77%

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-gepa/SKILL.md

Quality

Content

65%

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

The body is commendably lean and well-organized with a concrete core code pattern and useful guardrails. Its main gaps are an incomplete (placeholder-variable) code example, no validation/feedback checkpoint for an optimization workflow, and references to bundle files (API.md, axir-api.json, axir-capabilities.json, examples/) that are not present in the skill bundle.

Suggestions

Make the Core Pattern example copy-paste ready by defining or annotating reflection_client, request, and evaluator (e.g., show a minimal evaluator/construct call) to raise actionability.

Add an explicit validation/verification checkpoint for the optimize workflow (e.g., 'inspect result.pareto_front and metric_budget before accepting a candidate') so workflow_clarity is not capped at 2.

Either ship the referenced bundle files (API.md, axir-api.json, axir-capabilities.json, examples/) or remove/qualify the references so progressive disclosure points to real, resolvable material.

Dimension	Reasoning	Score
Conciseness	Quotes such as 'Language: Python.', 'Package: `axllm`.', and the compact 'Core Pattern' code block are lean and assume Claude's competence, with no padding or explanation of concepts Claude already knows. It is not 2 because there is no unnecessary explanation to tighten; every section earns its place, matching the 3-anchor 'lean and efficient'.	3 / 3
Actionability	Quotes 'engine = AxGEPA(reflection_client)' and 'result = engine.optimize(request, evaluator)' give a concrete code pattern, but reflection_client, request, and evaluator are undefined placeholders, so it is not copy-paste ready, matching the 2-anchor 'some concrete guidance but incomplete; missing key details'. It is not 3 because the example is not fully executable as written; not 1 because real API symbols and a runnable shape are provided rather than vague description.	2 / 3
Workflow Clarity	Quotes 'Start from package examples for exact native syntax before inventing a new call shape' give some sequencing, and the Core Pattern implies import-instantiate-optimize, but there is no validation or verification checkpoint for an optimization/batch-style task. This matches the 2-anchor 'steps listed but validation gaps'. It is not 3 because no explicit validate/feedback loop is present; not 1 because a basic sequence and guardrails exist.	2 / 3
Progressive Disclosure	Quotes 'Package API docs: `API.md` and `axir-api.json`', 'Capability manifest: `axir-capabilities.json`', and 'Runnable examples: `examples/`' signal one-level-deep references, but those files/directories are absent from the bundle (no references/, scripts/, assets/ dirs), so the navigation does not actually resolve. This matches the 2-anchor 'references present but not clearly signaled / could be better organized' in effect. It is not 3 because the referenced bundle files do not exist; not 1 because the body itself is well organized into clear sections rather than a monolithic wall.	2 / 3
	Total	9 / 12 Passed

Description

90%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is strong: it has an explicit 'Use when' trigger, a clear what/when pairing, good in-domain trigger terms, and a distinct niche tied to axllm/GEPA. Its only weakness is that the action language is a single generic verb ('writing') qualified by a list of areas rather than multiple distinct concrete actions.

Suggestions

Lead with concrete verbs (e.g., 'Run, inspect, and tune GEPA optimizations ...') instead of the single generic verb 'writing Python code' to lift specificity toward the top anchor.

Add one or two common phrasings a user might say (e.g., 'Pareto front', 'reflection loop') to broaden natural trigger coverage further.

Dimension	Reasoning	Score
Specificity	Quotes 'writing Python code with `axllm` for GEPA, Pareto tradeoffs, reflection clients, metric budgets, optimizer state, and artifacts' name the domain and many specific capability areas, but the only action verb is the generic 'writing', lacking multiple distinct concrete actions like the 3-anchor's 'extract/fill/merge'. It is not 3 because no multiple concrete verbs are listed; not 1 because the domain and sub-areas are specific rather than vague.	2 / 3
Completeness	Quotes 'Use when writing Python code with `axllm` for ...' provide an explicit 'Use when' trigger (when) plus a clear statement of what the skill does (writing Python code with axllm across the listed areas). It is not 2 because the trigger is explicit rather than implied; not below 3 because both what and when are clearly answered.	3 / 3
Trigger Term Quality	Quotes 'Python code', 'axllm', 'GEPA', 'Pareto tradeoffs', 'reflection clients', 'metric budgets', 'optimizer state', 'artifacts' give good coverage of the natural terms a user of this package would actually say. It is not 2 because it goes well beyond a single keyword with many relevant in-domain terms; these match the 3-anchor's 'good coverage of natural terms users would say'.	3 / 3
Distinctiveness Conflict Risk	Quotes 'with `axllm` for GEPA, Pareto tradeoffs, reflection clients' carve out a clear niche tied to a specific generated package, making accidental triggering of unrelated skills unlikely. It is not 2 because the triggers are distinctive rather than broadly overlapping; it matches the 3-anchor 'clear niche with distinct triggers'.	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-gepa/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.