cekura-metric-design

Use when the user asks to "create a metric", "write a metric", "design a metric", "build a metric for", "evaluate agent performance", "measure call quality", "track a KPI", "add a workflow metric", "improve my metric", "fix a metric", "debug metric results", "set up quality scoring", or "what metrics do I need". Also relevant when discussing LLM judge prompts, custom code metrics, evaluation triggers, VALID_SKIP patterns, section extraction, or metric best practices for Cekura voice AI agents. Covers both creating new metrics and reviewing, iterating on, or troubleshooting existing ones.

2.20x

Quality

90%

Does it follow best practices?

Impact

97%

2.20x

Average score across 2 eval scenarios

Securityby

Low

Low-risk findings worth noting

Quality

Content

81%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.

The body is a high-quality, action-oriented playbook with strong sequencing and feedback loops for a batch/cost-sensitive workflow. Its main weaknesses are mild verbosity in places and a dangling examples/ reference that breaks navigation.

Suggestions

Create the referenced examples/ files (llm-judge-metric.md, narrative-metric.md, custom-code-metric.py, section-extraction-metric.py) or remove those listings, since they are cited as deliverables but are absent from the bundle.

Tighten the 'Spirit vs Letter Principle' worked example and the Baseline Metrics table to their load-bearing points to recover token budget.

Add one fully inline copy-pasteable llm_judge metric example so the core action is executable without first opening references/prompt-patterns.md.

Dimension	Reasoning	Score
Conciseness	Dense, opinionated prose that largely assumes Claude's competence and avoids explaining basics, but a few sections (e.g., the spirit-vs-letter example and baseline-metrics table) could be tightened without losing value.	4 / 5
Actionability	Concrete named fields ('description', 'custom_code'), a copy-paste trigger prompt template, endpoint names, and eval-type tables give mostly executable guidance; full templates are deferred to reference files, leaving minor gaps inline.	4 / 5
Workflow Clarity	A clearly sequenced 6-step workflow with an explicit iterate/validate feedback loop ('Plan for at least one iteration'), a Cost Guard validation checkpoint before batch evaluation, and a Manual-Fix-First-Then-Labs recovery loop.	5 / 5
Progressive Disclosure	Overview is well-signaled and splits detail into real one-level reference files (prompt-patterns, pythonic-patterns, advanced-patterns, api-reference), but the body advertises an 'examples/' directory with four files that do not exist in the bundle.	4 / 5
	Total	17 / 20 Passed

Description

100%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.

The description is exemplary: it explicitly pairs 'what' and 'when', packs in natural user-synonym triggers, and scopes itself to a clear niche with minimal conflict risk. It is concise despite its breadth and uses appropriate third-person voice.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions ('create a metric', 'write a metric', 'design a metric', 'evaluate agent performance', 'measure call quality', 'track a KPI') plus concrete sub-tasks (LLM judge prompts, VALID_SKIP patterns, section extraction), giving comprehensive coverage.	5 / 5
Completeness	Explicit 'Use when...' clause with many concrete triggers, plus a clear 'what' statement ('Covers both creating new metrics and reviewing, iterating on, or troubleshooting existing ones').	5 / 5
Trigger Term Quality	Comprehensive natural trigger phrases with synonyms users would actually say — 'create/write/design/build a metric', 'improve my metric', 'fix a metric', 'debug metric results', 'what metrics do I need'.	5 / 5
Distinctiveness Conflict Risk	Narrow, well-scoped niche ('Cekura voice AI agents' metric design) with domain-specific triggers; the related eval-design skill is explicitly distinguished rather than blurred.	5 / 5
	Total	20 / 20 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: cekura-ai/cekura-skills
Path: cekura/skills/cekura-metric-design/SKILL.md
Commit: af2ceb6

Reviewed: 4 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.