ax-cpp-agent-optimize

Use when writing C++ 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

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./packages/cpp/skills/ax-cpp-agent-optimize/SKILL.md

Quality

Content

62%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 concise, well-sectioned overview with a distinct niche and accurate package facts, but it stops short of being actionable or workflow-shaped: the core code snippet is non-executable, and the optimization/verification workflow it gestures at is never sequenced or defined.

Suggestions

Expand the Core Pattern into one complete, executable example (includes, constructing a reflection client, an evaluator callback, and a request) drawn from the package examples, so the snippet is copy-paste ready.

Add a short sequenced workflow for a typical optimization run (start from example → define evaluator → run bounded optimize with explicit budget → verify playbook proposals pass the gate), including the steps of the "verification gate" that is currently referenced but never defined.

Make the referenced artifacts clearly navigable: link `API.md`, `axir-api.json`, `axir-capabilities.json`, and `examples/` with their actual paths so it is unambiguous where each lives.

Dimension	Reasoning	Score
Conciseness	The body is lean and factual — terse lines like "Real network support: yes" and "Scripted no-key transport support: yes" and a two-line core pattern with no padding or explanation of concepts Claude already knows. Every line carries package-specific information that is not inferable, matching the 'every token earns its place' anchor.	5 / 5
Actionability	The Core Pattern (`axllm::AxGEPA engine(reflection_client, options); auto result = engine.optimize(request, evaluator);`) is a real API shape but not executable — `reflection_client`, `options`, `request`, and `evaluator` are never defined or constructed, and no includes are shown. Guidance defers executability to external examples ("Start from package examples for exact native syntax"), which fits anchor 3 (incomplete, missing key details) better than anchor 4's mostly-executable standard.	3 / 5
Workflow Clarity	There is a rough implicit ordering in the guardrails (start from examples, use no-key examples for deterministic checks, treat AxIR as source of truth), but no sequenced workflow exists for any of the "When To Use" tasks — notably "keep only playbook proposals that pass the verification gate" references a verification gate that is never defined or given steps. This matches anchor 2 (rough sequence, many gaps, validation steps absent).	2 / 5
Progressive Disclosure	Good structure with clear sections and appropriately short body. References to `API.md`, `axir-api.json`, `axir-capabilities.json`, and `examples/` are present and one level deep, but they are named as bare facts rather than clearly signaled navigation (no links, and their location relative to the skill is ambiguous since no bundle directories exist), which is a minor organization gap matching anchor 4.	4 / 5
	Total	14 / 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: highly distinct, jargon-accurate triggers, explicit use-when guidance, and a specific domain. The main improvement is making the capability statement (what the skill does) explicit rather than folding it into the when-clause, and adding a couple of natural synonyms.

Suggestions

Lead with an explicit capability statement (e.g., "Writes and optimizes C++ agents with the `axllm` package...") before the "Use when..." trigger clause, so 'what' and 'when' are both clearly separated.

Add a few natural synonyms users might say, such as "few-shot", "AxAgent", or "prompt optimization", to broaden trigger coverage without adding vagueness.

Dimension	Reasoning	Score
Specificity	The description lists several concrete items — "agent optimization, verified agent-playbook evolution, evaluators, judges, optimizer artifacts, BootstrapFewShot, and GEPA" — anchored to a specific domain ("writing C++ code with `axllm`"). It sits above anchor 3 (which expects only 1-2 concrete actions) but below anchor 5, since most listed items are trigger contexts/objects rather than a comprehensive set of concrete actions.	4 / 5
Completeness	It has an explicit "Use when..." trigger clause and the what is present ("writing C++ code with `axllm`"), but the what is subordinate inside the when-clause rather than clearly stated as a standalone capability. This matches anchor 4 (both present, when could be more explicit) rather than anchor 5's fully separated what-and-when.	4 / 5
Trigger Term Quality	Good keyword coverage with natural ecosystem terms a user would say: "agent optimization", "evaluators", "judges", "BootstrapFewShot", "GEPA", "axllm", "C++". A few natural variations are missing (e.g., "few-shot", "AxAgent", "prompt optimization"), so it falls short of the comprehensive-synonym coverage of anchor 5.	4 / 5
Distinctiveness Conflict Risk	The combination of "C++", "axllm", "BootstrapFewShot", and "GEPA" carves out a clear niche with distinct triggers and minimal overlap risk with other skills — nothing generic remains that would fire the wrong skill.	5 / 5
	Total	17 / 20 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/cpp/skills/ax-cpp-agent-optimize/SKILL.md
Commit: 3f630a0

Reviewed: 3 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.