CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

ax-java-refine

Use when writing Java code with `dev.axllm:ax` 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
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

80%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, well-structured reference card: terse facts, real Java syntax, a clear API surface, and practical guardrails with no wasted tokens. Its weakness is workflow clarity — it never sequences the steps or explains when to use each of the four listed optimizers, leaving the agent to infer the path from Core Pattern and guardrails.

Suggestions

Add brief selection guidance for the optimizers listed under "Relevant API Surface" (e.g., when to use Ax.optimize vs Ax.playbook vs AxBootstrapFewShot vs AxGEPA), so the agent knows which entry point fits which refinement task.

Expand "Core Pattern" into a short numbered sequence: consult examples/ for the exact call shape, construct the evaluator, run the optimize call, then verify the result — turning the implicit flow into an explicit workflow with a verification step.

Include one complete runnable snippet (e.g., a no-key example) that defines reflectionClient, request, and evaluator, making the core pattern copy-paste ready.

DimensionReasoningScore

Conciseness

The body is lean throughout: a one-sentence intro, terse fact bullets ("Real network support: yes. Scripted no-key transport support: yes."), a two-line code pattern, and directive-style guardrails. No padding and no explanation of concepts Claude already knows; anti-drift against anchor 4 finds no over-explanation to trim.

5 / 5

Actionability

Provides concrete Java ("new AxGEPA(reflectionClient, java.util.Map.of())", "engine.optimize(request, evaluator)") and exact API names (AxGen, AxPlaybook, AxBootstrapFewShot, OptimizerEngine). Not a 5: reflectionClient, request, and evaluator are undefined, there is no fully runnable example, and the "Relevant API Surface" names symbols without showing how to use them — the "Start from package examples" guardrail partially mitigates this.

4 / 5

Workflow Clarity

The flow (When To Use → examples → Core Pattern → Guardrails) is present but implicit; there is no explicit sequence or validation checkpoint. Key decisions are left unresolved — when to choose Ax.optimize vs Ax.playbook vs AxBootstrapFewShot vs AxGEPA, and how to construct the evaluator and request. Anchor 3 ("sequence present but checkpoints missing or implicit") is the best fit; no destructive/batch operations, so the cap does not apply.

3 / 5

Progressive Disclosure

Under 50 lines with well-organized sections and no need for skill-bundle references (no references/, scripts/, or assets/ exist), which per the rubric's simple-skill note qualifies for 5. The one-level-deep pointers it does give ("Package API docs: `API.md` and `axir-api.json`", "Runnable examples: `examples/`") are clearly signaled under Package Facts.

5 / 5

Total

17

/

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: it pairs an explicit "Use when" trigger with four named capabilities and pins the niche to an exact package and language, keeping conflict risk minimal. The main gap is trigger-phrase breadth — it relies on package jargon rather than including the synonyms and intent-level phrases a user might naturally say.

DimensionReasoningScore

Specificity

Names the domain ("Java code with `dev.axllm:ax`") and four specific capabilities ("reward-scored generation", "iterative candidate improvement", "evaluator feedback", "optimizer-backed refinement patterns"). Not a 5 because "refinement patterns" is somewhat abstract and the description does not enumerate fully concrete operations the way the top anchor does.

4 / 5

Completeness

Explicitly answers both: when ("Use when writing Java code with `dev.axllm:ax`") and what ("reward-scored generation, iterative candidate improvement, evaluator feedback, and optimizer-backed refinement patterns"). The when is present and explicit but scoped to the package rather than intent-level trigger phrases, matching anchor 4 rather than 5.

4 / 5

Trigger Term Quality

Includes natural terms a user of this package would say: "Java code", "evaluator feedback", "optimizer", "refinement", "reward-scored". Coverage is good but missing common variations and synonyms (e.g., "improve output quality", "GEPA", "few-shot"), so it falls short of the comprehensive anchor 5.

4 / 5

Distinctiveness Conflict Risk

Names the exact package (`dev.axllm:ax`) and language (Java), creating a clear niche with minimal conflict risk; only sibling per-language skills for the same package could overlap, and "Java" differentiates them. Anti-drift check against anchor 4 confirms the exact package naming makes it more distinct than "mostly distinct".

5 / 5

Total

17

/

20

Passed

Validation

93%

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

CriteriaDescriptionResult

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/java/skills/ax-java-refine/SKILL.md

Reviewed

Table of Contents

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.