CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

ax-python-agent-optimize

Use when writing Python code with `axllm` for agent optimization, verified agent-playbook evolution, evaluators, judges, optimizer artifacts, BootstrapFewShot, and GEPA.

Quality

75%

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

Fix and improve this skill with Tessl

tessl review fix ./packages/python/skills/ax-python-agent-optimize/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

65%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.

The body is an admirably lean, fact-dense reference with genuinely non-obvious guardrails, but it functions more as a fact sheet than an operational guide. The core pattern is non-executable placeholder code, no step sequence exists for an optimization run, and the referenced package docs/examples are absent from the bundle.

Suggestions

Replace the placeholder core pattern with one complete, runnable example (real client construction, request, and evaluator) so the primary code snippet is copy-paste ready.

Add a short numbered workflow for a bounded optimization run (pick example -> adapt evaluator -> run with explicit budgets -> verify artifacts against baselineScore) with an explicit verification checkpoint.

Resolve the dangling references: either include `API.md`, `axir-api.json`, `axir-capabilities.json`, and `examples/` in the skill bundle with clear relative paths, or remove/repoint the citations.

DimensionReasoningScore

Conciseness

The ~45-line body is lean and every section carries package-specific, non-obvious knowledge (metric-call accounting, demo freshness rules, Pareto caveats) with no explanation of concepts Claude already knows. It matches anchor 5's "every token earns its place".

5 / 5

Actionability

The core pattern uses undefined placeholders (`reflection_client`, `request`, `evaluator`) making it a call shape rather than executable code, and the directive to "start from package examples" points to `examples/` files that are not present in the bundle. The concrete guardrails lift it above anchor 2 but it lacks anchor 4's mostly-executable guidance.

3 / 5

Workflow Clarity

"When To Use" and the guardrails convey goals, budgets, and validation concerns (verification gate, baseline reuse), but there is no sequenced procedure for an optimization run, so checkpoints remain implicit. This fits anchor 3's present-but-gapped sequencing better than anchor 2's rough-but-poorly-defined sequence.

3 / 5

Progressive Disclosure

Sections are well organized for a short skill, but the body cites `API.md`, `axir-api.json`, `axir-capabilities.json`, and `examples/` without paths, and none of these files exist in the skill bundle, leaving references present but unresolved. This matches anchor 3 (structure present, references not clearly signaled) rather than anchor 4's mostly-clear placement.

3 / 5

Total

14

/

20

Passed

Description

75%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, trigger-first description with a clear niche and specific named techniques. It pairs an explicit "Use when" clause with a concrete list of capability areas, only lacking the separate action-verb capability statement and synonym coverage of a top-scored example.

DimensionReasoningScore

Specificity

Lists several concrete capability areas ("agent optimization, verified agent-playbook evolution, evaluators, judges, optimizer artifacts, BootstrapFewShot, and GEPA"), but they are topic nouns rather than the concrete action verbs of anchor 5, and coverage has minor gaps.

4 / 5

Completeness

The explicit "Use when writing Python code with `axllm` for..." clause answers "when" with concrete triggers, and the "what" is conveyed inside that clause; it falls short of anchor 5, which states what and when as separate explicit elements.

4 / 5

Trigger Term Quality

"agent optimization", "evaluators", "judges", "BootstrapFewShot", and "GEPA" are natural terms a user of this package would say, but common synonyms like "few-shot" or "prompt optimization" are missing, matching the good-but-incomplete anchor 4.

4 / 5

Distinctiveness Conflict Risk

The named package (`axllm`) and techniques (GEPA, BootstrapFewShot) establish a clear niche with minimal conflict risk, but the generic terms "evaluators" and "judges" leave minor overlap risk with LLM-judging skills, fitting anchor 4 rather than 5.

4 / 5

Total

16

/

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/python/skills/ax-python-agent-optimize/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.