Skill Index — Progressive Disclosure Engine

Save 90%+ tokens. Agents scan the index first, load full skills only when executing. Inspired by Loki Mode's 3-Layer Progressive Disclosure memory architecture.

The Problem

Traditional approach: Load every SKILL.md to decide which skill to use.

• 30+ skills × ~3000 tokens each = 90,000+ tokens just for discovery
• Agent only needs ~500 tokens to make the choice
• 89,500 tokens wasted on reading full skills

The Solution: 3-Layer Loading

┌────────────────────────────────────────────────┐
│ LAYER 1: INDEX (~100 tokens per skill)          │
│ Always loaded. Quick scan. "What exists?"        │
│ Name + Domain + Triggers + 1-line summary       │
└─────────────────────┬──────────────────────────┘
                      │ (match found?)
                      v
┌────────────────────────────────────────────────┐
│ LAYER 2: SUMMARY (~300 tokens per skill)        │
│ Load when choosing. "Is this the right one?"     │
│ Description + When to use + Integration table   │
└─────────────────────┬──────────────────────────┘
                      │ (confirmed?)
                      v
┌────────────────────────────────────────────────┐
│ LAYER 3: FULL SKILL.md (1500-5000 tokens)       │
│ Load ONLY during execution. "How to use it."     │
│ Complete instructions, examples, red flags       │
└────────────────────────────────────────────────┘

Layer 1: Skill Index

Always loaded by agents at session start. Maximum 100 tokens per skill.

Engineering Swarm 🔧

Skill	Triggers	Summary
cm-tdd	test, TDD, red-green-refactor	Red-Green-Refactor cycle before any implementation
cm-debugging	bug, error, fix, debug, broken	Root cause analysis before fixing. 5-phase investigation
cm-quality-gate	deploy, ship, verify, quality, gate	6-gate verification: static analysis → blind review → ship
cm-test-gate	test setup, CI, test infrastructure	Setup 4-layer test gate for any project type
cm-code-review	review, PR, feedback, branch	Request reviews, handle feedback, complete branches
cm-codeintell	understand codebase, call graph, impact, architecture, what calls X, what breaks	AST knowledge graph + architecture diagrams + smart context (30% fewer tokens)
cm-reactor	rewrite, pivot, migrate, direction change, architecture wrong, restructure	TRIZ-powered strategic codebase re-direction — 5-phase incremental migration
cm-clean-code	clean, dead code, refactor, code smell, DRY, SOLID, naming, messy	7-point hygiene gate: dead code, imports, naming, SRP, DRY, nesting

Operations Swarm ⚙️

Skill	Triggers	Summary
cm-safe-deploy	deploy, staging, production, release	Multi-gate deploy pipeline with rollback strategy
cm-identity-guard	git push, deploy, Cloudflare, Supabase, account	Verify identity before any push/deploy to prevent wrong-account
cm-git-worktrees	branch, isolate, worktree, parallel	Isolated git worktrees for feature work
cm-terminal	command, terminal, run, execute	Safe terminal execution with progress logging

Security Swarm 🔒

Skill	Triggers	Summary
cm-secret-shield	secret, token, leak, security, scan, pre-commit, gitleaks, rotate	Defense-in-depth: pre-commit hooks, repo scanning, token lifecycle

Product Swarm 🎨

Skill	Triggers	Summary
cm-brainstorm-idea	analyze, enhance, improve, initiative, evaluate, review product	Strategic analysis gate: 9 Windows + Double Diamond → 2-3 qualified options
cm-planning	plan, design, brainstorm, feature	Brainstorm intent → write implementation plan → then code
cm-ux-master	UI, UX, design, interface, usability	48 UX Laws + 37 Design Tests + Figma/Stitch integration
cm-ui-preview	preview, visual, mockup, render UI, wireframe	Live UI concept generation via Google Stitch or Pencil MCP
cm-dockit	docs, documentation, knowledge base, SOP	Generate complete knowledge base from codebase
cm-project-bootstrap	new project, init, bootstrap, setup	Full project setup: design system → staging → CI → deploy
cm-jtbd	customer discovery, JTBD, jobs to be done, product-market fit, why users churn	JTBD canvas: Switch Interview → Outcome Metrics → Opportunity Scoring

Growth Swarm 📈

Skill	Triggers	Summary
cm-content-factory	content, blog, article, marketing	AI content engine: research → generate → audit → deploy
cm-ads-tracker	pixel, tracking, GTM, Meta, TikTok, Google Ads	Complete conversion tracking setup across platforms
cm-cro-methodology	conversion, A/B test, landing page, funnel	CRO audit: funnel mapping → persuasion → objection handling

Orchestration Swarm 🎯

Skill	Triggers	Summary
cm-execution	execute, implement, plan, RARV, batch	Execute plans: batch mode, subagent-driven, parallel, or RARV
cm-continuity	memory, context, CONTINUITY, learnings	Working memory protocol: read at start, update at end
cm-skill-mastery	which skill, skill list, discover, help	Meta-skill: when to invoke skills, how to create new ones
cm-safe-i18n	translate, i18n, language, localize	Safe translation with multi-pass batching and audit gates
cm-skill-chain	chain, pipeline, workflow, multi-step, full process	Compose skills into automated multi-step pipelines
cm-deep-search	semantic search, find docs, large codebase, qmd	Optional power-up for semantic search across large projects
cm-readit	audio, TTS, read aloud, voice, speech, SpeechSynthesis, MP3 player	Web audio engine — TTS reader, pre-recorded MP3 player, Voice CRO triggers
cm-how-it-work	how does X work, explain, architecture overview	Explain how a system, feature, or codebase works

Workflow Commands 🖥️

Skill	Triggers	Summary
cm-start	/cm-start, start workflow, begin objective, new task	Orchestrate full cm-* workflow from objective to production code
cm-dashboard	/cm-dashboard, kanban, task board, show tasks, status board	Render Kanban board from cm-tasks.json — visual task overview
cm-status	/cm-status, quick status, progress, what's next, blocked	Ultra-concise progress summary: done / next / blocked

Planned (not yet implemented) 🧪

Skills below do not have SKILL.md files yet.

Skill	Triggers	Summary
mom-test (planned)	customer interview, validate idea, leading questions	Customer interviews without leading questions
release-it (planned)	production, circuit breaker, timeout, chaos	Build production-ready systems with stability patterns
medical-research (planned)	medical, OB/GYN, clinical, evidence-based	Evidence-based medical writing with citation standards
tailwind-mastery (planned)	Tailwind, utility-first, responsive, v4	Tailwind CSS utilities, responsive, accessibility
pandasai-analytics (planned)	DataFrame, analytics, forecast, time-series	Natural language DataFrame analysis and visual reports
cm-google-form (planned)	Google Form, sheet, App Script	Form-to-sheet integration with auto-retry

Community Skills Registry 🌐

59 pre-indexed community skills from skills.sh. Install on-demand: npx skills add {source} --skill {name} -a antigravity If tech not listed here → fall back to npx skills find "{keyword}"

🖥️ Frontend Frameworks (10)

Skill	Source	Install Keywords
react-best-practices	vercel-labs/agent-skills	react, react performance, RSC
composition-patterns	vercel-labs/agent-skills	react composition, compound components
react-native-guidelines	vercel-labs/agent-skills	react native, expo, mobile app
react-native-best-practices	callstackincubator/agent-skills	react native, expo, mobile
nextjs-app-router-patterns	wshobson/agents	next.js, app router, RSC
react:components	google-labs-code/stitch-skills	react components, stitch
cm-ux-master	UI/UX, design system, frontend	UI/UX design intelligence + Harvester v4
angular-directives	analogjs/angular-skills	angular, directives, components
svelte-best-practices	community	svelte, sveltekit
astro-best-practices	community	astro, static site, islands

🎨 CSS & UI Design (8)

Skill	Source	Install Keywords
web-design-guidelines	vercel-labs/agent-skills	UI review, accessibility, UX audit
tailwind-design-system	wshobson/agents	tailwind, design tokens, utility CSS
tailwind-v4-shadcn	jezweb/claude-skills	tailwind v4, shadcn/ui
tailwindcss-advanced-layouts	josiahsiegel/claude-plugin-marketplace	tailwind layouts, responsive
remotion-best-practices	remotion-dev/skills	remotion, programmatic video
shadcn-ui-patterns	community	shadcn, radix, component library
css-architecture	community	CSS modules, BEM, CSS-in-JS
framer-motion	community	animation, framer motion, gestures

🗄️ Databases & ORM (9)

Skill	Source	Install Keywords
supabase-postgres-best-practices	supabase/agent-skills	supabase, postgres, RLS, SQL
prisma-database-setup	prisma/skills	prisma, schema, migration
prisma-client-api	prisma/skills	prisma client, type-safe queries
prisma-expert	sickn33/antigravity-awesome-skills	prisma, advanced patterns
drizzle-orm	community	drizzle, ORM, type-safe SQL
mongodb-patterns	community	mongodb, mongoose, NoSQL
pinecone-db	community	pinecone, vector DB, embeddings
turso-database	community	turso, libSQL, edge database
planetscale	community	planetscale, MySQL, branching

🔐 Authentication (6)

Skill	Source	Install Keywords
clerk-nextjs-patterns	clerk/skills	clerk, auth, nextjs auth
clerk-auth-v5	clerk/skills	clerk v5, middleware, session
better-auth-best-practices	community	better-auth, auth library
next-auth-v5	the-misfits/agent-skills	nextauth, auth.js, OAuth
auth-js-v5	the-misfits/agent-skills	auth.js, session, JWT
supabase-auth	community	supabase auth, RLS, magic link

💳 Payments & Commerce (4)

Skill	Source	Install Keywords
stripe-client	community	stripe, payment, checkout
stripe-webhook	community	stripe webhook, event handling
stripe-checkout	community	stripe checkout, subscription
stripe-connect	community	stripe connect, marketplace

🧪 Testing (5)

Skill	Source	Install Keywords
testing-library	community	testing library, react testing
vitest-best-practices	the-misfits/agent-skills	vitest, unit test, coverage
playwright-e2e	the-misfits/agent-skills	playwright, e2e, browser test
cypress-patterns	community	cypress, e2e, component test
testing-guidelines	anthropics/skills	test strategy, TDD, test design

☁️ Cloud & Deployment (7)

Skill	Source	Install Keywords
vercel-deploy-claimable	vercel-labs/agent-skills	vercel deploy, preview URL
vercel-deployment-best-practices	community	vercel, edge functions, ISR
azure-deployment	microsoft/skills	azure, cloud, Microsoft
aws-deployment	microsoft/skills	AWS, Lambda, S3
docker-expert	community	docker, container, Dockerfile
docker-compose-orchestration	community	docker compose, multi-container
multi-stage-dockerfile	community	docker build, multi-stage, CI

🤖 AI & ML (5)

Skill	Source	Install Keywords
vercel-ai-sdk-best-practices	community	AI SDK, streaming, LLM
langchain-skills	community	langchain, agents, RAG
openai-skills	community	openai, GPT, embeddings
tavily-search	community	tavily, web search, research
ai-sdk-patterns	community	AI SDK, tool calling, generative UI

📡 APIs & Integrations (3)

Skill	Source	Install Keywords
nextjs-supabase-auth	sickn33/antigravity-awesome-skills	next.js + supabase, full-stack
supabase-edge-functions	nice-wolf-studio/claude-code-supabase-skills	edge functions, Deno
nextjs-seo	laguagu/claude-code-nextjs-skills	SEO, meta tags, sitemap

📱 Mobile & Cross-Platform (2)

Skill	Source	Install Keywords
expo-patterns	community	expo, EAS, mobile deploy
capacitor-patterns	community	capacitor, ionic, hybrid app

---

Discovery Protocol:

1. Check THIS index first (59 skills above)
   → Match found? → npx skills add {source} --skill {name} -a antigravity
2. No match? → npx skills find "{keyword}" on skills.sh
3. Still nothing? → Proceed without skill, note gap in .cm-skills-log.json

Trigger points:
  - cm-start Step 0.5 (before planning)
  - cm-planning Phase A Step 4 (scope analysis)
  - cm-execution Pre-flight (before dispatch)
  - cm-execution RARV (after 2 failed retries)

Layer 2: Skill Summary

Loaded when an agent has matched a skill from Layer 1 and needs more context.

To access Layer 2, read the first 20 lines of the corresponding SKILL.md (frontmatter + first section).

Example: cm-tdd Layer 2
  - Full name: "TDD — Test Before Write"
  - When to Use: ALWAYS before implementing any feature or bugfix
  - Process: Red (write failing test) → Green (minimal code) → Refactor (clean up)
  - Integration: Works with cm-execution (subagent TDD), cm-quality-gate (coverage check)
  - Token cost: ~4200 tokens for full skill

Layer 3: Full SKILL.md

Loaded ONLY when executing. Read the complete SKILL.md file.

# Agent loads full skill only when ready to execute
view_file /path/to/skills/cm-tdd/SKILL.md

Usage Protocol

1. AT SESSION START:
   - Load this skill index (Layer 1) — costs ~2500 tokens total
   - Now you know what 60 skills do without reading any of them

2. WHEN MATCHING A TASK:
   - Scan Layer 1 triggers to find matching skill(s)
   - If unsure, read Layer 2 (first 20 lines of SKILL.md)
   - Choose the best skill

3. WHEN EXECUTING:
   - Load Layer 3 (full SKILL.md) for the chosen skill
   - Follow the skill's instructions completely

4. SAVINGS:
    - Without index: 34 × 3000 = 102,000 tokens
    - With index: 2800 + 300 + 3000 = 6,100 tokens
    - Saved: ~95,900 tokens (~94%)

Skill Domain Mapping

For dynamic agent selection (Phase 3), skills map to domains:

engineering: [cm-tdd, cm-debugging, cm-quality-gate, cm-test-gate, cm-code-review, cm-codeintell, cm-reactor, cm-clean-code]
operations: [cm-safe-deploy, cm-identity-guard, cm-git-worktrees, cm-terminal, cm-safe-i18n]
security: [cm-secret-shield]
product: [cm-brainstorm-idea, cm-planning, cm-ux-master, cm-ui-preview, cm-dockit, cm-project-bootstrap, cm-jtbd]
growth: [cm-content-factory, cm-ads-tracker, cm-cro-methodology]
orchestration: [cm-execution, cm-continuity, cm-skill-mastery, cm-skill-chain, cm-deep-search, cm-readit, cm-how-it-work]
commands: [cm-start, cm-dashboard, cm-status]
# planned (no SKILL.md yet): mom-test, release-it, medical-research, tailwind-mastery, pandasai-analytics, cm-google-form

The Bottom Line

Know everything. Load nothing. Execute precisely. Save 90%+ tokens.