omni-inference

The core OpenAI-compatible inference endpoints: chat completions, embeddings, images, audio (TTS/STT), moderations, rerank, and the Responses API. The primary integration surface for AI agents.

Quality

70%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Advisory

Suggest reviewing before use

Fix and improve this skill with Tessl

tessl review fix ./skills/omni-inference/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 genuinely actionable API reference with strong executable examples, error mappings, and model-discovery guidance in its custom sections, but it is monolithic and redundant — overlapping auto-generated and custom sections plus many empty curl stubs inflate the token budget without progressive disclosure into local files. Workflow clarity is adequate for a reference doc but lacks explicit checkpoints.

Suggestions

Collapse the redundant auto-generated Endpoints stubs (empty `-d '{}'` bodies) into a compact endpoint table, keeping the rich examples only in the custom sections.

Split per-capability detail (Image, TTS/STT, Web search/fetch) into reference files under references/ and have SKILL.md point to them, enabling one-level-deep progressive disclosure.

Add a short top-of-body workflow (set OMNIROUTE_URL/KEY → discover models via /v1/models → call endpoint → handle error code) so the usage sequence is explicit rather than implicit per section.

Dimension	Reasoning	Score
Conciseness	The custom section is lean, but the auto-generated Endpoints section repeats the frontmatter description in the Overview and emits ~20 near-identical curl stubs with empty `-d '{}'` bodies plus repeated auth headers, and chat/images/audio/embeddings are documented twice (auto-gen vs custom); could be tightened, yet it does not explain concepts Claude already knows, so above level 1.	2 / 3
Actionability	The custom sections provide copy-paste-ready, fully executable curl commands with real JSON bodies (model, messages, stream), discover commands (`curl .../v1/models \| jq`), response shapes, and error→action mappings; this matches the executable anchor, while the empty auto-gen stubs are supplementary rather than the primary guidance.	3 / 3
Workflow Clarity	Each capability follows a discover→example→errors pattern, but there is no overarching multi-step workflow with explicit validation checkpoints and the body is a flat catalog rather than a sequenced process; it is clearer than level 1 but lacks the checkpoints of level 3.	2 / 3
Progressive Disclosure	The body is well-sectioned by capability and points to an external OpenAPI spec (`docs/openapi.yaml`) and an entry-point SKILL URL, but no local bundle/reference files exist and all ~585 lines live inline in one monolithic file, so content that could be split stays inline; the under-50-line exemption does not apply.	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 specific, well-keyworded catalog of the inference surface, but it lacks any explicit "Use when" trigger guidance and its broad scope risks overlapping with narrower sibling skills. Strong on what it does, weaker on when Claude should select it.

Suggestions

Add an explicit trigger clause, e.g. "Use when integrating an AI agent with OpenAI/Anthropic-compatible chat, embeddings, images, audio, moderation, or rerank endpoints."

Narrow the distinctiveness by stating what this skill is NOT (e.g., provider configuration/setup), since the body delegates setup to a separate entry-point skill.

Lead with the verb/action form of capabilities ("Generate chat completions, transcribe audio, embed text") to raise specificity beyond a noun list.

Dimension	Reasoning	Score
Specificity	The description enumerates concrete capabilities — "chat completions, embeddings, images, audio (TTS/STT), moderations, rerank, and the Responses API" — a comprehensive, specific list rather than vague language; not the level-2 single-domain case.	3 / 3
Completeness	It clearly answers "what" but offers no "Use when..." or explicit trigger guidance — "The primary integration surface for AI agents" only implies the "when," so per the rubric it is capped at 2 rather than the full level-3 what-and-when.	2 / 3
Trigger Term Quality	Terms like "chat completions," "embeddings," "TTS/STT," "rerank," "Responses API," and "OpenAI-compatible" are natural vocabulary a developer would actually say when needing this surface, giving good coverage; not the partial-variation level 2.	3 / 3
Distinctiveness Conflict Risk	The very broad scope (chat, images, audio, embeddings, moderation, rerank, web) overlaps with narrower sibling skills aggregated in the body, so it could still trigger for the wrong skill; not the clear-niche level 3 but more specific than level 1.	2 / 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
skill_md_line_count	SKILL.md is long (590 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: diegosouzapw/OmniRoute
Path: skills/omni-inference/SKILL.md
Commit: 3070925

Reviewed: 2 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.