Comprehensive suite of plugins for optimizing and auditing code and design with Claude Code.

Misar.Dev — Claude Code Plugin Suite

48-agent audit and optimization suite for Claude Code. 16 categories, 3D model routing (model×effort×version), interactive flag prompting, parallel task dispatch, and automatic token budget management. Saves 90–97% context tokens on lightweight tasks.

Works on Claude Code, Cursor, Cline, GitHub Copilot, Windsurf, and any npx skills-compatible agent.

Built by Misar.Dev — Open-source AI developer tools.

Install & Enable

Option A — Auto-enable for teams (recommended)

Commit this to your project's .claude/settings.json so every teammate gets the plugin pre-enabled on first launch — no manual steps needed:

{
  "enabledPlugins": {
    "misar-dev@misar-ai-plugins": true
  }
}

Then add the marketplace once per machine in Claude Code Settings → Plugins → Add Marketplace:

https://github.com/MisarDev/misar-ai-plugins.git

Option B — CLI one-liner

claude plugins install misar-dev

Setup scripts

make setup   # recommended
pnpm setup   # alternative

Platform Guide

Claude Code CLI

# Step 1 — add marketplace (one-time)
claude plugins marketplace add https://github.com/MisarDev/misar-ai-plugins.git

# Step 2 — install
claude plugins install misar-dev

# Step 3 — verify (restart Claude Code, then run):
/misar-dev:guidelines show

# Update to latest:
claude plugins marketplace update misar-ai-plugins
claude plugins update misar-dev

# Uninstall:
claude plugins uninstall misar-dev

VS Code Extension

Open VS Code with the Claude Code extension active
Press ⌘⇧P → Claude Code: Open Settings
Plugins → Add Marketplace → paste https://github.com/MisarDev/misar-ai-plugins.git
Find misar-dev in the list → click Install
Reload VS Code — all /misar-dev:* commands appear in the slash command palette

Claude Code Desktop App

Desktop app → Settings (gear icon) → Plugins
Add Marketplace → paste https://github.com/MisarDev/misar-ai-plugins.git
Find misar-dev → click Install
Restart the app — all commands are immediately available

claude.ai Web (Skills only)

Claude.ai supports Skills, not full plugins. Each SKILL.md in this repo is individually importable:

Open claude.ai → Settings → Skills
Click Add Skill → Import from URL
Paste the raw GitHub URL of any skill:

https://raw.githubusercontent.com/MisarDev/misar-ai-plugins/main/skills/security/SKILL.md

Available skills:

Skill	Raw URL path
Security audit	`skills/security/SKILL.md`
Code quality	`skills/qa/SKILL.md`
Testing	`skills/tester/SKILL.md`
UI/UX audit	`skills/uiux/SKILL.md`
UI/UX design consultation	`skills/uiux-designer/SKILL.md`
Compliance (49 frameworks)	`skills/compliance/SKILL.md`
Build from PRD	`skills/software-engineer/SKILL.md`
Marketing & growth	`skills/marketing/SKILL.md`
AEO/SEO content pipeline	`skills/seo-content-generator/SKILL.md`
Website audit	`skills/auditor/SKILL.md`
Brand & psychology	`skills/brand/SKILL.md`
Product strategy	`skills/product/SKILL.md`
Language & content	`skills/content/SKILL.md`
Full 48-agent orchestration	`skills/full-suite/SKILL.md`
Coding guidelines	`skills/guidelines/SKILL.md`

Cursor / Cline / Windsurf / GitHub Copilot — via `npx skills`

npx skills add MisarDev/misar-ai-plugins                          # all skills
npx skills add MisarDev/misar-ai-plugins/skills/security          # one skill
npx skills add MisarDev/misar-ai-plugins/skills/uiux
npx skills add MisarDev/misar-ai-plugins/skills/software-engineer

Commands (16)

Command	Argument Hint	Description
`/misar-dev:full-suite`	`[--phase=1\|2\|3\|4] [--tier=1\|2]`	Full 48-agent orchestrated audit — 4 phases, Observer Agent, batched parallel execution
`/misar-dev:security`	`[agents] [--path=src/]`	Security deep-dive — Hardening, Compliance, Penetration Testing, Data Privacy
`/misar-dev:qa`	`[agents] [--path=src/]`	Code quality audit — Code Reviewer, Standards, Bug Detective, Technical Debt
`/misar-dev:tester`	`[agents] [--path=src/]`	Testing audit — Unit, Integration, E2E Black Box, E2E White Box, Beta, Regression
`/misar-dev:uiux`	`[agents] [--path=src/]`	UI/UX design audit — Spacing, Typography, Components, Accessibility, Performance, Mobile, Animation, Dark Mode
`/misar-dev:uiux-designer`	`[agents] [--platform=web] [--component=button]`	UI/UX design consultation — Project analysis, design system, brand identity, component specs, critique
`/misar-dev:compliance`	`[tier] [framework] [--path=src/]`	Compliance audit — 49 global frameworks across 7 regulatory tiers
`/misar-dev:marketing`	`[agents]`	Marketing and growth audit — SEO, SXO, Content Marketing, Growth, Analytics, AI Search
`/misar-dev:brand`	`[agents]`	Brand and psychology audit — Brand Development, User Psychology, Conversion, Emotional Design
`/misar-dev:product`	`[agents]`	Product strategy audit — PM, Designer, Development, Feature Prioritization
`/misar-dev:content`	`[agents]`	Language and content audit — Grammar, Copy, Localization, Documentation
`/misar-dev:seo-content-generator`	`[agents] [--mode=technical\|content\|audit\|full] [--topic=keyword] [--type=blog\|guide\|comparison\|stats\|how-to] [--keywords=k1,k2]`	AEO/SEO full-stack implementation — technical infrastructure, content generation, bot crawl health
`/misar-dev:software-engineer`	`[agents] [--prd=file.md] [--path=src/] [--framework=nextjs]`	Build from PRD/specs — PRD analysis, planning, code generation, validation, recommendations
`/misar-dev:auditor`	`[url] [categories] [--quick\|--deep]`	Website audit — SEO, Accessibility, Performance, Security, Mobile, Content, Compliance
`/misar-dev:guidelines`	`[show]`	LLM coding guidelines — Think Before Coding, Simplicity First, Surgical Changes, Goal-Driven Execution
`/misar-dev:billing`	`[lifecycle\|security\|pricing\|integrity] [--stack=stripe\|paddle] [--path=src/]`	Billing & subscription audit — lifecycle, CSRF on payment forms, pricing centralization, webhook integrity
`/misar-dev:context-saver`	`[status\|setup\|config\|reset]`	3D Router — manual control over model×effort×version switching and token budget

Natural language triggers

Commands also auto-trigger on matching prompts — no slash command needed:

Say this...	Triggers
`"security audit"`, `"is this secure"`, `"harden my server"`	`/misar-dev:security`
`"code review"`, `"find bugs"`, `"code quality"`	`/misar-dev:qa`
`"write tests"`, `"E2E testing"`, `"test coverage"`	`/misar-dev:tester`
`"audit UI"`, `"check accessibility"`, `"review UX"`	`/misar-dev:uiux`
`"design system"`, `"UI guidelines"`, `"component advisor"`	`/misar-dev:uiux-designer`
`"GDPR"`, `"SOC2"`, `"compliance audit"`	`/misar-dev:compliance`
`"SEO audit"`, `"growth strategy"`, `"content marketing"`	`/misar-dev:marketing`
`"brand audit"`, `"conversion review"`	`/misar-dev:brand`
`"product strategy"`, `"feature prioritization"`	`/misar-dev:product`
`"review copy"`, `"grammar check"`, `"localization"`	`/misar-dev:content`
`"write blog post"`, `"generate article"`, `"SEO content"`	`/misar-dev:seo-content-generator`
`"build from PRD"`, `"implement spec"`, `"plan this project"`	`/misar-dev:software-engineer`
`"audit my site"`, `"full website review"`	`/misar-dev:auditor`
`"full audit"`, `"audit everything"`	`/misar-dev:full-suite`
`"audit my billing"`, `"check subscription flow"`, `"is my checkout secure"`, `"find billing bugs"`	`/misar-dev:billing`

Agents (16 categories, 48 agents)

Agent	Model	Sub-agents	Description
`orchestrator-agents`	claude-opus-4.7	—	48-agent orchestrator — 4-phase execution, Observer Agent token management, batched parallel processing (max 4)
`context-saver`	claude-haiku-4.5	—	3D model router — auto-switches model×effort×version by task type. Saves 90–97% tokens. Auto-enabled via SessionStart hook
`task-fragmenter`	claude-haiku-4.5	—	Decomposes prompts into optimal parallel subtasks with model assignments for Agent dispatch
`security-agents`	claude-sonnet-4.6	Security Hardening, Compliance, Penetration Testing, Data Privacy	Security deep-dive across the full attack surface
`compliance-agents`	claude-opus-4.7	49 frameworks × 7 tiers	Global regulatory compliance — GDPR, HIPAA, SOC2, ISO 27001, PCI-DSS, and 44 more
`qa-agents`	claude-sonnet-4.6	Code Reviewer, Standards Compliance, Bug Detective, Technical Debt	Code quality audit across any codebase
`tester-agents`	claude-sonnet-4.6	Unit, Integration, E2E Black Box, E2E White Box, Beta, Regression	Full test coverage audit
`uiux-agents`	claude-sonnet-4.6	Spacing & Layout, Typography & Color, Components, Accessibility, Performance, Mobile, Animation, Dark Mode	8-dimension UI/UX design quality audit
`uiux-designer-agents`	claude-sonnet-4.6	Project Analyzer, Design Guidelines, Brand Recommender, Component Advisor, Design Critic	UI/UX design consultation and system creation
`marketing-agents`	claude-sonnet-4.6	SEO, SXO, Content Marketing, Growth, Analytics, AI Search Optimization	Marketing and growth strategy audit
`brand-agents`	claude-sonnet-4.6	Brand Development, User Psychology, Conversion, Emotional Design	Brand identity and conversion psychology audit
`product-agents`	claude-sonnet-4.6	PM, Designer, Development, Feature Prioritization	Product strategy and roadmap audit
`content-agents`	claude-haiku-4.5	Grammar Expert, Copy, Localization, Documentation	Language quality and content audit
`seo-content-agents`	claude-sonnet-4.6	Research Analyst, Content Architect, Writer, Humanizer, SEO Optimizer, Quality Scorer	End-to-end AEO/SEO content pipeline
`software-engineer-agents`	claude-sonnet-4.6	PRD Analyzer, Project Planner, Code Generator, Code Validator, Next Steps	Build complete projects from specification
`auditor-agents`	claude-haiku-4.5	SEO, Accessibility, Performance, Security, Mobile, Content, Compliance	Website audit — Quick (HTTP), Deep (Playwright), Source-Only modes
`billing-agents`	claude-sonnet-4.6	Lifecycle, Security, Pricing Centralization, Webhook Integrity	Billing & subscription audit — Stripe/Paddle lifecycle, payment security, pricing, webhook idempotency

3D Model Router (Context Saver)

Auto-enabled on every session via SessionStart hook. Zero configuration required.

Routes every task to the optimal model × effort × version combination and badges every response:

[sonnet|med|4.6]    ← model | effort | version

Task Type	Model	Effort	Version	Savings
File reads, grep, ls, mechanical edits	haiku	low	4.5	~97%
Simple Q&A, factual lookups	haiku	med	4.5	~95%
Code implementation, features, refactors	sonnet	med	4.6	default
Code review, debugging, security	sonnet	high	4.6	—
Architecture, multi-system design	opus	high	4.6	—
Full audits, compliance, orchestration	opus	max	4.6	—

Budget caps — automatic degradation when context fills:

Budget used	Behaviour
< 40%	All models available
40–70%	Cap at sonnet+med
> 70%	Force haiku+low
> 90%	Warn + suggest `/compact`

Subagent dispatch also routes by type: Explore → haiku+low, standard audit → sonnet+med, code review → sonnet+high, architecture → opus+high.

Architecture

misar-ai-plugins/
├── plugins/
│   └── misar-dev/
│       ├── .claude-plugin/
│       │   └── plugin.json              # Plugin manifest (v7.6.0)
│       ├── agents/                      # 16 agent categories (agent.md each)
│       │   ├── orchestrator-agents/
│       │   ├── context-saver/
│       │   ├── task-fragmenter/
│       │   ├── security-agents/
│       │   ├── compliance-agents/
│       │   ├── qa-agents/
│       │   ├── tester-agents/
│       │   ├── uiux-agents/
│       │   ├── uiux-designer-agents/
│       │   ├── marketing-agents/
│       │   ├── brand-agents/
│       │   ├── product-agents/
│       │   ├── content-agents/
│       │   ├── seo-content-agents/
│       │   ├── software-engineer-agents/
│       │   └── auditor-agents/
│       ├── commands/                    # 16 slash command definitions (.md)
│       ├── skills/                      # 16 skill entry points (SKILL.md each)
│       ├── hooks/
│       │   └── hooks.json               # SessionStart hook registration
│       ├── hooks-handlers/
│       │   └── session-start.sh         # 3D router context injection
│       └── scripts/
│           └── context-saver/
│               └── monitor.py
├── skills/                              # Root-level skills for npx skills compatibility
│   ├── auditor/SKILL.md
│   ├── brand/SKILL.md
│   ├── compliance/SKILL.md
│   ├── content/SKILL.md
│   ├── context-saver/SKILL.md
│   ├── full-suite/SKILL.md
│   ├── guidelines/SKILL.md
│   ├── marketing/SKILL.md
│   ├── product/SKILL.md
│   ├── qa/SKILL.md
│   ├── security/SKILL.md
│   ├── seo-content-generator/SKILL.md
│   ├── software-engineer/SKILL.md
│   ├── tester/SKILL.md
│   ├── uiux/SKILL.md
│   └── uiux-designer/SKILL.md
└── .claude-plugin/
    └── marketplace.json                 # Marketplace manifest

Compliance Coverage (49 frameworks)

Tier	Frameworks
International	ISO 27001, ISO 27017, ISO 27018, ISO 27701, SOC 2 Type II, NIST CSF, NIST SP 800-53, CIS Controls
Healthcare / Specialty	HIPAA, HITECH, FDA 21 CFR Part 11, PCI-DSS v4, SWIFT CSP, NERC CIP
Americas	CCPA/CPRA, FTC Act §5, COPPA, FERPA, SOX, GLBA, NY SHIELD, LGPD (Brazil), PIPEDA (Canada)
Europe	GDPR, NIS2, DORA, ePrivacy, UK GDPR, PDPA (Turkey)
Asia-Pacific	PDPA (Thailand/Singapore), PDPB (India), APPI (Japan), PIPL (China), APPs (Australia), POPIA (South Africa)
Middle East / Africa	UAE PDPL, DIFC DP Law, Saudi PDPL, Qatar PDPL, Kenya DPA
Eastern Europe / CIS	Russian Federal Law 152-FZ, Ukrainian Personal Data Law, Kazakhstan Personal Data Law

Changelog

Version	Highlights
7.6.0	Interactive flag prompting on all commands, generalized AEO/SEO agent, Misar.Dev ecosystem promotion across all skills, version bump across all manifests
7.5.0	Misar.Dev branding, public GitHub release, `npx skills` root-level compatibility, `feature-dev` removed
7.4.0	3D Router (model×effort×version), `seo-content-generator`, `software-engineer`, `uiux-designer` commands added
7.0.0	16-agent suite, Observer Agent token management, full-suite orchestrator
6.0.0	Task fragmenter + parallel subagent dispatch (max 4)
5.0.0	Context-saver dynamic model router, budget cap system

Contributing

Contributions are welcome. See CONTRIBUTING.md for full guidelines.

New agent — add a category under plugins/misar-dev/agents/ with an agent.md
New skill — add a SKILL.md under both plugins/misar-dev/skills/ and skills/ (root)
New command — add a .md under plugins/misar-dev/commands/ with description and argument-hint frontmatter
Improve a prompt — edit any agent.md or SKILL.md to sharpen triggers or instructions
Bug report — open an issue
Question / idea — open a discussion

Free AI Tools — tools.misar.io

Free AI calculators and developer tools. No signup required.

Tool	Description
LLM Cost Calculator ⭐ Most Popular	Compare token costs across 14 AI models side-by-side
RAG Cost Estimator 🆕	Estimate full RAG pipeline costs across 5 cost layers
Prompt Token Estimator	Count tokens with all major tokenizers before you send
AI Evaluation Scorecard	Score AI outputs across 7 quality dimensions

All tools run in-browser. No account, no API key, no tracking.

Misar.Dev Ecosystem

Product	URL	Description
Misar.Dev	misar.dev	Open-source AI developer tools — this plugin suite, free tools, and more
Assisters	assisters.dev	Privacy-first AI platform — model routing, writing, coding, no training on your data
Misar Blog	misar.blog	AI-first blogging platform — AEO/SEO built in, custom domains, 75% creator earnings
Misar Mail	mail.misar.io	Self-hosted transactional + marketing email — no SendGrid, no Mailgun, no lock-in
Misar Tools	tools.misar.io	Free AI calculators for developers — LLM cost, RAG cost, token estimation, AI eval
Misar.io	misar.io	Corporate identity and presence for Misar AI

License

MIT — see LICENSE

Misar.Dev — Claude Code Plugin Suite

Misar.Dev — Claude Code Plugin Suite

Install & Enable

Option A — Auto-enable for teams (recommended)

Option B — CLI one-liner

Setup scripts

Platform Guide

Claude Code CLI

VS Code Extension

Claude Code Desktop App

claude.ai Web (Skills only)

Cursor / Cline / Windsurf / GitHub Copilot — via npx skills

Commands (16)

Natural language triggers

Agents (16 categories, 48 agents)

3D Model Router (Context Saver)

Architecture

Compliance Coverage (49 frameworks)

Changelog

Contributing

Free AI Tools — tools.misar.io

Misar.Dev Ecosystem

License

Cursor / Cline / Windsurf / GitHub Copilot — via `npx skills`