CtrlK
BlogDocsLog inGet started
Tessl Logo

ax-cpp-ai

Use when writing C++ code with `axllm` for named deployment profiles, generic provider clients, model selection, OpenAI-compatible calls, Responses, Gemini, Anthropic, routers, and balancers.

55

Quality

69%

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/cpp/skills/ax-cpp-ai/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

50%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 dense, information-rich reference-style body that scores well on precision (exact defaults, error classes, and naming conventions) but functions as an inlined API reference rather than a SKILL.md overview. The absence of any bundle files, near-total lack of executable C++ examples, and untrimmed cross-language detail hold all dimensions to the midpoint.

Suggestions

Split reference-grade detail (per-model sampling rules, retry backoff parameters, Astra session adapter internals, the 'Relevant API Surface' symbol list) into a reference file such as API.md or a provider-rules.md, keeping SKILL.md as a concise overview that links to them.

Add one or two complete, copy-paste-ready C++ examples in the body — e.g., a client with base URL + credential provider, and a basic routed/balanced setup — since only a two-line core pattern currently exists.

Add short step sequences for the most common flows (choose profile → configure credentials → make call → route/failover), and trim the Python/Go/Java/Rust method-name listings that are irrelevant to a C++-only skill.

DimensionReasoningScore

Conciseness

The body avoids explaining concepts Claude already knows and almost every line carries package-specific facts, but ~160 lines include reference-grade minutiae (per-model sampling rules, retry backoff formulas, five-language adapter parity, a ~40-symbol API listing) and cross-language API name lists (Python/Go/Java/Rust) that are trimmable from a C++ skill. Mostly efficient material that could be significantly tightened or offloaded.

3 / 5

Actionability

Rules are exact (error classes like `AxAIServiceTimeoutError`, retry defaults '500, 408, 429, 502, 503, 504 and 529', concrete signatures like `add_child_agent(namespace, name, child)`), but the only executable code is a two-line core-pattern snippet; a code-generation skill with no complete runnable example is a bigger gap than the anchor-4 'minor gaps' allows.

3 / 5

Workflow Clarity

Orientation exists (a 'When To Use' bullet list, Guardrails, and a sequenced hint like 'Start with `examples/adaptive_balancer_no_key` for store/reducer syntax, then use the cataloged provider-backed adaptive-balancer example'), but common flows — profile selection → credentials → call → routing — are never explicitly sequenced and validation checkpoints are absent (no destructive operations, so the workflow cap does not apply).

3 / 5

Progressive Disclosure

Clear section headers and well-signaled pointers to `API.md`, `axir-capabilities.json`, and `examples/`, but no bundle files actually exist in the skill directory, and ~160 lines of reference-level detail (the material API.md would hold) are inlined in SKILL.md. This matches the anchor of content that should live in a separate file being inline behind decent structure.

3 / 5

Total

12

/

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 solid description with an explicit 'Use when' trigger and a concrete list of capability areas anchored by a specific package name and language. Its main gaps are the absence of an action-oriented statement of what the skill provides and a few missing natural synonyms for covered features.

DimensionReasoningScore

Specificity

Lists several specific capability areas ('named deployment profiles, generic provider clients, model selection, OpenAI-compatible calls, Responses, Gemini, Anthropic, routers, and balancers') but presents them as nouns rather than actions and omits body topics like structured output, retries, and prompt caching. Matches the anchor for several specific items with minor coverage gaps, not comprehensive enough for a 5.

4 / 5

Completeness

The 'when' is explicit and strong ('Use when writing C++ code with `axllm` for...'), and the 'what' is conveyed as a list of capability areas. Not a 5 because the description never states what the skill actually provides (e.g., API reference and usage rules) — the 'what' is only implied by the task list.

4 / 5

Trigger Term Quality

Includes natural terms a user would say ('C++ code', 'axllm', 'OpenAI-compatible', 'Gemini', 'Anthropic', 'routers', 'balancers'). Good coverage but missing common variations like 'chat completions', 'failover', or the other provider names (DeepSeek, Grok, Groq) the body covers.

4 / 5

Distinctiveness Conflict Risk

'C++ code with `axllm`' carves a clear, distinct niche unlikely to trigger the wrong skill. Minor overlap risk remains with sibling skills for the same package in other languages, since provider names (OpenAI, Gemini, Anthropic) are shared.

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
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.