ax-cpp-gepa

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

Quality

70%

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-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.

A compact, well-organized reference-style body with a clean API surface and useful guardrails, but the core code example is not executable as written and it references documentation/example files that are not bundled. It also lacks an explicit validation loop for the optimization workflow.

Suggestions

Either bundle the referenced files (API.md, axir-api.json, axir-capabilities.json, examples/) or remove/relabel the references so navigation does not point to missing material.

Make the Core Pattern example executable by defining the placeholder variables (reflection_client, options, request, evaluator) or by linking directly to a complete runnable example in examples/.

Add a short validation checkpoint for the optimization workflow (e.g., how to inspect the Pareto front, confirm metric-budget consumption, and retry on failed candidates).

Dimension	Reasoning	Score
Conciseness	The body is lean: tight bullet lists, a short code shape, and no padding or explanation of concepts Claude already knows, so every section earns its tokens.	3 / 3
Actionability	It gives a concrete API symbol list and a call shape ("axllm::AxGEPA engine(reflection_client, options); auto result = engine.optimize(request, evaluator);"), but the snippet uses undefined placeholders (reflection_client, options, request, evaluator) and is not copy-paste executable, matching the incomplete-guidance anchor at 2 rather than the fully-executable anchor at 3.	2 / 3
Workflow Clarity	The Guardrails supply loose procedural direction (start from examples, choose provider-api vs no-key, treat AxIR as source of truth), but there is no sequenced multi-step workflow and no validation/feedback checkpoint for a batch optimization operation, which the rubric caps at 2.	2 / 3
Progressive Disclosure	Sections are well organized and references are signaled in Package Facts ("API.md", "axir-api.json", "axir-capabilities.json", "examples/"), but none of those referenced files actually exist in the bundle, so the one-level-deep navigation points to missing material rather than a 3-worthy navigable structure.	2 / 3
	Total	9 / 12 Passed

Description

75%

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 concise, well-targeted description that clearly states both purpose and trigger conditions with a strong, distinctive niche. Its main weakness is jargon-heavy trigger terms and topic-noun phrasing rather than concrete user-facing actions.

Suggestions

Reframe the capability list around concrete actions a user would request (e.g., "run GEPA optimization", "build Pareto fronts", "track metric budgets and reflection calls") rather than topic nouns.

Soften jargon with natural variations users might actually say (e.g., pair "reflection clients" with "self-reflection prompts", "Pareto tradeoffs" with "Pareto front").

Dimension	Reasoning	Score
Specificity	The phrase "writing C++ code with `axllm` for GEPA, Pareto tradeoffs, reflection clients, metric budgets, optimizer state, and artifacts" names the domain and several specific capability areas, but these are topics/nouns rather than concrete verbs (e.g., "extract text", "fill forms"), so it does not reach the multiple-concrete-actions bar of a 3.	2 / 3
Completeness	It opens with an explicit "Use when writing C++ code with `axllm` for ..." clause that states both what the skill does (write C++ with axllm for GEPA-related tasks) and when to use it, satisfying the explicit-trigger requirement that unlocks a 3.	3 / 3
Trigger Term Quality	It includes some recognizable trigger terms ("writing C++ code", "axllm", "GEPA", "Pareto"), but the list leans on technical jargon ("reflection clients", "metric budgets", "optimizer state") with no common user-facing variations, falling short of the natural-term coverage expected at 3.	2 / 3
Distinctiveness Conflict Risk	The combination of C++ + `axllm` + GEPA/Pareto/reflection/optimizer-state is a narrow, specialized niche with distinct triggers, making it unlikely to fire for an unrelated skill.	3 / 3
	Total	10 / 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/cpp/skills/ax-cpp-gepa/SKILL.md
Commit: 9087a01

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.