By Ovid
Conduct multi-agent audits for architecture, code quality, accessibility across web/mobile/desktop, security flaws, and spec alignment; enforce TDD guardrails on fixes, review git branches before merge, standardize Makefiles, and generate verified reports with fixes.
Comprehensive multi-agent accessibility audit of user-facing code — supports web, mobile (iOS/Android/React Native/Flutter), desktop, CLI, and games — dispatches specialists for screen readers, vision, motor, cognitive, and multimedia concerns, verifies findings, and produces an actionable report with WCAG 2.2 AA/AAA ratings
Multi-agent architecture analysis — dispatches specialists for structure, coupling, integration, error handling, and security, verifies findings, and produces a comprehensive report of strengths and flaws with evidence
Use when reviewing current branch for bugs before pushing or merging, when wanting a thorough multi-agent review of local changes, or when preparing work for human review
Check that requirements, designs, and implementation plans are aligned — finds coverage gaps, scope creep, and design mismatches, then rewrites tasks in TDD red/green/refactor format
Guided fixing of architectural flaws from an agentic-architecture report — validates findings, writes tests, applies fixes with developer approval, and tracks status in the report
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Good developers don't ship code without tests, reviews, CI/CD, QA, and incident response. Layer after layer of defense-in-depth, because no one gets it right the first time.
AI coding assistants don't have any of that. They get it wrong and ship it anyway. PAAD adds the defense-in-depth.
Most AI coding assistants have the same four problems. PAAD addresses each one. See also, this horrendously long explanation.
| Problem | What goes wrong | Status |
|---|---|---|
| Pushback | AI is a smarmy little git. Push back? No. Push to production? Absolutely. Push problems to Future You? Every time. | Largely solved |
| Alignment | AI hears "add a button" and delivers a button, a modal, a notification system, and a config page. The button doesn't work. | Largely solved |
| Architecture | Your architecture is a house of cards. AI's solution? More cards. Taller. | Partially solved |
| Discipline | PAA works — when you actually use it. Skip pushback, ignore alignment, defer the review, and every problem comes back worse. The process requires discipline to deliver results. | It's on you |
PAAD is a system of AI agent skills that gives your coding assistant the tools to catch these problems before they compound. It supports Claude Code natively, with experimental support for Cursor, Kiro, and Antigravity.
WARNING: PAAD is brutally honest. It will tell you when your spec is flawed, your plan is misaligned, your architecture has problems, or your code has bugs. If you don't want to hear that, don't install PAAD.
Also, while this methodology has served me well, it eats tokens like popcorn at a horror movie. I'm not aiming for quantity, I'm aiming for quality.
First, you might forget commands. That's OK. Just run /paad:help to see all available skills and their usage.
Before I start working in any repo, I run /paad:makefile to set up a standard Makefile with targets for building, testing, linting, and formatting. This gives me a consistent interface for common tasks and encourages good habits. More importantly, the code coverage, linting, and formatting tools are critical to help PAAD generate production-quality code and catch issues early.
Then I follow this general workflow:
/paad:pushback to get a critical review of your spec./paad:alignment to check that your implementation plan matches your requirements and design./paad:agentic-review to catch bugs, security holes, and integration problems before merging.I often run /paad:pushback and /paad:alignment more than once to catch issues the first run (or introduced).
Also ...
/paad:agentic-architecture to catch structural problems before they compound./paad:agentic-a11y to catch accessibility barriers before they go live./paad:vibe to get TDD guardrails and contextual suggestions./plugin marketplace add Ovid/paad
/plugin install paad@paad
Add to your project's .claude/settings.json so teammates are automatically prompted:
{
"extraKnownMarketplaces": {
"paad": {
"source": {
"source": "github",
"repo": "Ovid/paad"
}
}
}
}
You can invoke skills explicitly with /paad:pushback, /paad:alignment, etc. But Claude Code also recognizes skill names in natural language — saying "I want you to pushback on this design" or "run an architecture review" will trigger the corresponding skill. The /paad: slash command forces invocation when you want to be certain the skill runs.
PAAD skills use the same SKILL.md format that Cursor skills expect.
All skills (bash/zsh):
cp -r kiro_and_antigravity/skills/.kiro/skills/* .cursor/skills/
All skills (Windows):
xcopy kiro_and_antigravity\skills\.kiro\skills\* .cursor\skills\ /E /I
One skill (e.g., pushback):
cp -r kiro_and_antigravity/skills/.kiro/skills/pushback .cursor/skills/
All skills (bash/zsh):
cp -r kiro_and_antigravity/skills/.kiro/skills/* .kiro/skills/
All skills (Windows):
xcopy kiro_and_antigravity\skills\.kiro\skills\* .kiro\skills\ /E /I
npx claudepluginhub ovid/paad --plugin paad[Beta] 6つの専門エージェントチームでコードベースのアーキテクチャを多角的に分析するスキル
AI-powered development workflow automation - Phase-based planning, implementation orchestration, preflight code quality checks with security scanning, ship-it workflow, and development principles generator for CLAUDE.md
Code transformation: Dev SDLC orchestrator (code-shipping pipeline), plan, assert, audit, review, test, refactor, debug, for-sure. Hosts engineering agents.
Configurable code review with radical candor - choose between harsh direct feedback or constructive caring criticism. Features Technical.md for project standards, architectural context analysis, actionable fixes, decision register for tracking review decisions, seamless todo integration, and live browser QA via Chrome.
Persona-driven AI development team: orchestrator, team agents, review agents, skills, slash commands, and advisory hooks for Claude Code
Helder's personal SDLC toolbelt for AI coding agents — from PRD to ship. Bundles the tracer-bullet workflow alongside TDD, code review, audits, and shipping skills.