cli-eval

Create and run evaluation suites, watch live benchmark progress, view scorecards, compare model performance, and integrate eval runs with CI workflows from the CLI.

Quality

77%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No known issues

Fix and improve this skill with Tessl

tessl review fix ./skills/cli-eval/SKILL.md

Quality

Content

72%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is highly actionable and well-organized with copy-paste commands, but it carries redundant generated reference content and lacks explicit validation checkpoints while presenting two divergent command syntaxes.

Suggestions

Resolve the divergent command syntax: the generated section uses 'eval run <suiteId>' / 'eval list' while the curated section uses 'eval suites run <suiteId>' / 'eval list' — pick one and remove the duplicate 'eval list' and 'eval get' headers.

Tighten conciseness by collapsing the generated per-subcommand 'Example' blocks (which omit their own flags) into the curated sections, or showing each example with its flags inline.

Add an explicit validation/feedback checkpoint for eval runs (e.g. after 'eval suites run', poll 'eval get <runId>' until status is 'completed' before reading 'eval results'/'eval scorecard') rather than burying it in the Errors section.

Dimension	Reasoning	Score
Conciseness	Mostly efficient command-driven content, but the generated subcommand reference is redundant with the curated section and its 'Example' blocks repeat bare commands without the flags they just listed; not 3 due to this redundancy, not 1 because it avoids explaining concepts Claude already knows.	2 / 3
Actionability	Provides fully executable commands with real flags and realistic model IDs (e.g. 'omniroute eval suites run <suiteId> --model claude-sonnet-4-6 --watch'), JSONL sample format, jq pipelines, and a copy-paste CI threshold snippet; matches the copy-paste-ready anchor.	3 / 3
Workflow Clarity	Steps are listed and roughly sequenced (create suite -> run -> watch/poll -> results/scorecard) with a polling hint in Errors, but validation checkpoints are only implicit and two competing command syntaxes ('eval run <suiteId>' vs 'eval suites run <suiteId>') create ambiguity; not 3 due to missing explicit validate->fix->retry checkpoints, not 1 because a sequence is present.	2 / 3
Progressive Disclosure	A single self-contained SKILL.md with no bundle files and one clearly-signaled one-level external link (the CLI entry-point skill), organized into clean topical sections; per the rubric's simple-skill note, well-organized sections with no external references qualify for 3.	3 / 3
	Total	10 / 12 Passed

Description

82%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is specific, action-rich, and well-scoped to a distinct niche, but it omits an explicit 'Use when...' trigger clause, which caps completeness at 2.

Suggestions

Add an explicit 'Use when...' clause naming natural triggers, e.g. 'Use when running model evaluations, benchmarking models, checking scorecards, or gating CI on eval scores.'

Include a few more everyday user phrasings ('benchmark models', 'eval models', 'CI score gate') alongside the technical terms to broaden trigger coverage.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions ('Create and run evaluation suites', 'watch live benchmark progress', 'view scorecards', 'compare model performance', 'integrate eval runs with CI workflows'), matching the anchor for specific concrete actions rather than the partial-coverage anchor below.	3 / 3
Completeness	Clearly answers 'what' with concrete actions but lacks any 'Use when...' clause or equivalent explicit trigger guidance, which the rubric caps at 2; not 3 due to the missing explicit 'when', not 1 because 'what' is strong.	2 / 3
Trigger Term Quality	Uses natural domain terms a user would say ('evaluation suites', 'benchmark progress', 'scorecards', 'compare model performance', 'CI workflows'), giving good coverage of the niche's vocabulary; not score 2 because these are not generic jargon but terms users actually invoke.	3 / 3
Distinctiveness Conflict Risk	Scoped to the CLI eval/benchmark/scorecard/CI niche with distinct triggers, unlikely to fire for unrelated skills; matches the clear-niche anchor rather than the overlap anchor.	3 / 3
	Total	11 / 12 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: diegosouzapw/OmniRoute
Path: skills/cli-eval/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.