skills.rest
levnikolaevich
levnikolaevichCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

ln-61-skill-reviewer

Reviews standalone skills and their distribution surfaces for publication readiness.

556|83|Updated Oct 24, 2025
Search Tags:#quality gate#Marketplace#skill review#behavioral testing#release readiness#frontmatter validation#Claude code
One-click install
npx skills add https://github.com/levnikolaevich/claude-code-skills --skill ln-61-skill-reviewer
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ln-61-skill-reviewer
Source: https://github.com/levnikolaevich/claude-code-skills/tree/main/plugins/maintainer-suite/skills/ln-61-skill-reviewer
Command: npx skills add https://github.com/levnikolaevich/claude-code-skills --skill ln-61-skill-reviewer

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Publishing an AI skill without a rigorous review risks broken installation, wrong triggering, unsafe behavior, or invalid contracts reaching users. This Skill provides a structured, evidence-based review process that validates skill structure, frontmatter, host catalog integration, and real behavioral activation before release.

Core Features & Use Cases

  • Structural Validation: Checks frontmatter naming, description quality, canonical layout, and repository-defined contracts against each configured host catalog.
  • Behavioral Review: Derives positive and close-negative prompts, walks normal and failure scenarios, and uses fresh independent contexts to verify activation boundaries.
  • Evidence-Based Verdicts: Tracks every checklist item as PROVEN, CLEARED, or UNPROVEN and issues a PASS, PASS WITH CONCERNS, FAIL, or BLOCKED verdict with cited findings.
  • Use Case: Before publishing a new skill to a Claude Code or Codex marketplace, run this review to confirm the skill triggers correctly, stays standalone, and passes all host-native validators.

Quick Start

Review the skill in plugins/my-suite/skills/my-skill for publication readiness and report a verdict with evidence.

Frequently Asked Questions about ln-61-skill-reviewer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I review a Claude Code skill before publishing it?▼

Run a structured review that validates frontmatter naming and description quality, checks the canonical skill layout, runs repository and host-native validators, and tests behavioral activation with positive and close-negative prompts. The review ends with a PASS, PASS WITH CONCERNS, FAIL, or BLOCKED verdict.

What makes a skill description pass review?▼

A passing description states the capability, a positive trigger, and an important near-negative boundary while staying within the host's length limit. It must avoid claims broader than what the workflow actually supports.

Does this review modify the repository or publish anything?▼

No. The review is strictly read-only: it does not edit files, stage changes, publish packages, update marketplaces, or create GitHub state. Any diagnostic artifact must be temporary and disclosed in the residual-risk section.

What happens when a required validator tool is unavailable?▼

Tool absence is not itself a defect. The review falls back to manual YAML, naming, path, and catalog checks, and only returns BLOCKED when the missing capability prevents a reliable verdict.

When should a skill review return a FAIL verdict?▼

FAIL applies when a confirmed defect can cause wrong triggering, unsafe behavior, broken installation, or an invalid contract. Critical failures must be reproduced before assigning a blocking severity.

Related Skills

Explore More Software Engineering Skills→
hidonguyenhidonguyen

ck:skill-creator

Automate creation, evaluation, and packaging of Claude skills with SKILL.md validation.

Community
Advanced
anselsourceanselsource

write-plugin-marketplaces

Create and distribute Claude Code plugin marketplaces with marketplace.json.

Official
Intermediate
sohei56sohei56
15

codebase-audit

Audit the accumulated repository at HEAD for cross-cutting specification, logic, redundancy, documentation, and security defects.

Community
Advanced
mouratomourato

quality-assurance

Automate code quality assurance with testing, linting, and verification commands.

Community
Advanced