Name: ax-agent-optimize
Rating: 65.60000000000001 (1 reviews)
Author: ax-llm

ax-agent-optimize

This skill helps an LLM generate correct AxAgent tuning and evaluation code using @ax-llm/ax. Use when the user asks about agent.optimize(...), judgeOptions, eval datasets, optimization targets, saved optimizedProgram artifacts, or agent optimization guidance.

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 ./website/static/typescript/.well-known/agent-skills/ax-agent-optimize/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.

A highly actionable reference skill with strong executable code patterns, undermined by cross-section redundancy and a monolithic single-file structure that inlines reference material rather than splitting it into bundle files. Workflow sequencing is implicit rather than a crisp validated pipeline.

Suggestions

Dedupe the metric-vs-judge and Plain AxGen guidance that repeats across 'Use These Defaults', 'Metric vs Judge', and 'Dataset And Judge Rules' into a single authoritative section to recover token budget.

Move the dense reference material (Eval Semantics, Delegation Optimization Notes, Dataset And Judge Rules) into references/ files linked one level deep from SKILL.md to improve progressive disclosure.

Add a short numbered optimization workflow (choose shape → run optimize → save artifact → replay on held-out task → inspect any zero-score task) with explicit verification checkpoints to lift workflow clarity.

Dimension	Reasoning	Score
Conciseness	The body is domain-specific and mostly assumes Claude's competence, but guidance recurs across sections (e.g. 'a custom metric overrides the built-in judge path' appears in both 'Metric vs Judge' and 'Dataset And Judge Rules'), so it could be tightened. It is not a 1 because it avoids explaining generic concepts, and not a 3 because the cross-section redundancy adds tokens that do not earn their place.	2 / 3
Actionability	Provides multiple complete, copy-paste-ready TypeScript patterns (Canonical, Minimal, Deterministic Metric, Built-In Judge, Plain AxGen) plus concrete API rules with specific field names, so the guidance is fully executable.	3 / 3
Workflow Clarity	A decision guide and save/load/compare sequence are present with a debugging feedback loop ('inspect that task first instead of adding more optimizer rounds'), but the overall optimize→save→verify flow is presented as scattered reference notes rather than a clearly numbered sequence with explicit validation checkpoints.	2 / 3
Progressive Disclosure	The body is well-sectioned with headers and links to external example files, but no bundle files exist and substantial reference material (Eval Semantics, Delegation Notes, Dataset rules) is inlined in SKILL.md rather than split into one-level-deep reference files.	2 / 3
	Total	9 / 12 Passed

Description

92%

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-structured description with an explicit trigger clause, concrete actions, and natural-plus-API keywords. Its main weakness is modest overlap risk with the sibling ax-gepa/ax-playbook optimization skills due to the generic 'agent optimization guidance' phrasing.

Suggestions

Tighten the 'agent optimization guidance' trigger to make clear it is the AxAgent-specific path (e.g. 'Use when the user wants to tune an AxAgent via agent.optimize(...)') to reduce overlap with the ax-gepa and ax-playbook skills.

Dimension	Reasoning	Score
Specificity	Names multiple concrete actions — 'generate correct AxAgent tuning and evaluation code using @ax-llm/ax' — pinning the framework and two specific code-generation tasks rather than vague language.	3 / 3
Completeness	Explicitly answers both 'what' ('helps an LLM generate correct AxAgent tuning and evaluation code') and 'when' with an explicit 'Use when the user asks about ...' trigger clause.	3 / 3
Trigger Term Quality	Combines natural phrasings ('agent optimization guidance', 'optimization targets', 'eval datasets') with the API terms a developer working with this library would actually say ('agent.optimize(...)', 'judgeOptions'), giving good coverage.	3 / 3
Distinctiveness Conflict Risk	The AxAgent + agent.optimize(...) anchoring carves a niche, but 'agent optimization guidance' is broad and the skill family (ax-gepa for generic optimize, ax-playbook) creates real overlap risk with sibling optimization skills.	2 / 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: website/static/typescript/.well-known/agent-skills/ax-agent-optimize/SKILL.md
Commit: 7911e27

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.