Help us improve
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
By Cheggin
Deterministic phase-gated skill enforcement for AI coding agents. Blocks file edits until earlier workflow phases complete.
npx claudepluginhub cheggin/skill-chain --plugin skill-chainRun automated WCAG 2.1 AA accessibility audits using axe-core via Playwright. Use when implementing or reviewing UI features to enforce accessibility compliance, catch color contrast failures, validate keyboard navigation, check screen reader landmarks, verify alt text and ARIA attributes, or gate feature completion on accessibility standards.
Adapt designs to work across different screen sizes, devices, contexts, or platforms. Implements breakpoints, fluid layouts, and touch targets. Use when the user mentions responsive design, mobile layouts, breakpoints, viewport adaptation, or cross-device compatibility.
Create new agent definitions with specific instruction sets, skill assignments, and behavioral rules. Use when the user wants a new type of agent (e.g., research-papers agent, customer-support agent, data-pipeline agent) or wants to modify an existing agent's behavior.
Clean AI-generated code slop with a regression-safe, deletion-first workflow and optional reviewer-only mode
Integrate PostHog product analytics into a Next.js application with page views, event tracking, conversion funnels, A/B testing via feature flags, and API access for programmatic dashboard queries. Use when setting up product analytics, tracking user actions, defining conversion funnels, or enabling A/B experiments.
Modifies files
Hook triggers on file write and edit operations
Share bugs, ideas, or general feedback.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Harness-native ECC plugin for engineering teams - 63 agents, 251 skills, 79 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Reliable automation, in-depth debugging, and performance analysis in Chrome using Chrome DevTools and Puppeteer
Autonomous startup builder — idea to deployed company. 121 skills (37 OMC + 84 custom), 32 agents, deterministic skill-chain enforcement across 4 flows, hooks on all 11 Claude Code event types, bundled OMC runtime + MCP server + bridge CLI. Standalone: no dependency on the oh-my-claudecode plugin. Validated: built 3 startups (B2C, devtool, B2B SaaS) from single sentences.
Autonomous AI agent that builds UIs in a loop
Deterministic phase-gated skill enforcement for AI coding agents. Forces agents to follow a defined workflow order — no skipping discovery, no jumping straight to code.
Works with Claude Code and OpenAI Codex CLI. Zero dependencies.
AI coding agents skip steps. Ask one to "build a website" and it jumps straight to writing code — skipping research, design, brand guidelines, accessibility, and quality checks. Prompt instructions help but agents forget ~20% of them. You need mechanical enforcement.
Define a skill chain — an ordered sequence of phases that must complete before the agent can edit files:
{
"flows": {
"website-design": {
"trigger_skill": "website-creation",
"gate_patterns": ["app/**", "components/**", "*.tsx", "*.css"],
"phases": [
{ "name": "discovery", "required": ["shape", "brand-guidelines"] },
{ "name": "foundation", "required": ["website-creation", "impeccable"] },
{ "name": "refinement", "anyOf": { "of": ["layout", "typeset"], "min": 1 } },
{ "name": "quality", "required": ["polish"] },
{ "name": "ship", "required": ["deploy-pipeline"] }
]
}
}
}
When the agent tries to edit app/page.tsx without having invoked shape and brand-guidelines first:
[skill-chain] Flow "website-design", phase 1 "discovery" is incomplete.
Invoke before continuing: shape, brand-guidelines.
The edit is blocked at the tool level. Not a suggestion — a hard gate.
Edit/Write calltrigger_skill was most recently called)gate_patternsexit 2 = edit blocked with a message naming the missing skillsNo state files. No external services. The session transcript is the single source of truth.
/plugin marketplace add Cheggin/skill-chain
/plugin install skill-chain@skill-chain
This registers the enforcer hook and all 8 example skills automatically. Skills become available as /skill-chain:<skill-name>.
git clone https://github.com/Cheggin/skill-chain
cp skill-chain/skill-chain.mjs your-project/
cp skill-chain/skill-chains.json your-project/
cp -r skill-chain/skills/ your-project/skills/
cd your-project
node skill-chain.mjs # or use setup.mjs to wire hooks automatically
node setup.mjs # auto-detects runtime
node setup.mjs --runtime claude # Claude Code only
node setup.mjs --runtime codex # Codex CLI only
This adds the enforcer as a PreToolUse hook in .claude/settings.json or .codex/hooks.json.
Edit skill-chains.json to define your flows. Each flow needs:
| Field | Purpose |
|---|---|
trigger_skill | Which skill activation starts this flow |
gate_patterns | Glob patterns for files that are gated |
phases | Ordered list of phases with required skills |
Skills are just markdown files at skills/<name>/SKILL.md with YAML frontmatter. The enforcer only checks skill names (slugs) — it never reads the skill files themselves.
This repo includes 8 example skills for the website-design chain. Use them as-is or write your own.
Three ways to define what "complete" means for a phase:
// ALL must fire
{ "name": "discovery", "required": ["shape", "brand-guidelines"] }
// At least ONE must fire
{ "name": "styling", "oneOf": ["tailwind-setup", "css-modules-setup"] }
// At least N of M must fire
{ "name": "refinement", "anyOf": { "of": ["layout", "typeset", "colorize", "adapt"], "min": 2 } }
The included skill-chains.json defines a 5-phase website build flow:
| Phase | Skills | What it enforces |
|---|---|---|
| Discovery | shape, brand-guidelines | Research UX and brand before touching code |
| Foundation | website-creation, impeccable | Set up the project with design principles |
| Refinement | 1+ of layout, typeset | Apply at least one design refinement pass |
| Quality | polish | Final quality pass before shipping |
| Ship | deploy-pipeline | Structured deploy process |
User: "Build me a landing page"
Agent: *immediately writes app/page.tsx*
Result: Generic layout, no brand consistency, no responsive design, ships broken