PAAD plugin
Skills for Claude Code and other agents: spec pushback, six-specialist branch review, architecture and accessibility audits
- 153 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Skills start several subagents and read the whole branch
- /fix-architecture, /vibe and /backlog fix edit project files
- Reports and configuration are written into project folders
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/paadDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Skills start several subagents and read the whole branch
- /fix-architecture, /vibe and /backlog fix edit project files
- Reports and configuration are written into project folders
This is third-party code. Review the repository files before installing.
What it does
PAAD (Pushback, Alignment, Architecture, Discipline) is for people who write code with AI and want to stay the engineer: its skills review each stage while it can still change. The main set: /pushback argues with your spec before anyone builds it, /alignment checks that the plan and the spec say the same thing, /agentic-review sends a branch to six specialist reviewers before merge, /agentic-architecture finds structural debt with five specialists, /fix-architecture works through the findings one at a time with tests first, and /agentic-a11y audits accessibility against WCAG 2.2 AA. There are lighter skills too: /vibe for small fixes with TDD guardrails, /rethink checks the premises behind a recommendation, /test-roadmap builds a suite that catches real regressions, /handoff hands the session to a fresh one in writing, /makefile creates a Makefile. Experimental ones: /agentic-dedup finds duplicated meaning, /agentic-owasp checks code against OWASP Top 10:2025, /backlog clears fixed bugs from the backlog. Every skill can be tuned per project through paad/config/paad.md and paad/config/skill-name.md. Security findings stay out of the reports you commit.
Who it is for. Developers who run a project with an AI agent and want critique at the spec, plan, architecture and branch level.
Good fit when
- A spec needs a check for weak points before the build
- A branch is ready to merge and deserves several independent reviewers
- A project has built up structural debt that must be worked through in order
- A session with the agent must be handed to a fresh one as a written note
Not a fit when
- You need one quick comment on code: a six-agent review is heavier
- You need a formal security audit: the OWASP skill is marked experimental
- Team policy forbids running several subagents in one task
Example request
Run /pushback on docs/spec.md and list the weak points before we start writing codeLimitations
The agentic skills launch several subagents and use a noticeable number of tokens. Some skills are experimental, and so is configuration through the paad/config files. In the Claude desktop app the /plugin marketplace add command only opens Settings, so add the plugin through Settings, Plugins. Read the six-agent review results yourself: some findings will be debatable, and decisions on changes stay with you.
How to disable. In Claude Code remove the plugin with /plugin; for other agents run npx skills remove by skill name. Delete paad/ folders with reports and configuration in the project by hand.
FAQ
How do I install PAAD in Cursor or Codex?
With npx skills@latest add Ovid/paad, which installs the set into more than 70 agents including Cursor, Codex, GitHub Copilot, Gemini CLI, Cline, Zed, Warp, Amp and OpenCode. Claude Code has its own plugin marketplace.
What does PAAD stand for?
Pushback, Alignment, Architecture and the Discipline to actually run them. It rhymes with pad.
What is new in 2.0.0?
Per-project skill configuration through config files, and security findings now stay out of committed reports. Details are in the paad--v2.0.0 release notes and the changelog.
How do I tune a skill for my project?
Drop instructions into paad/config/paad.md for all skills or paad/config/skill-name.md for one. The skill reads the file on invocation. The feature is experimental.
What does /agentic-review do?
Six specialists review your branch before merge. You can pass a base branch and a path: /agentic-review [base-branch] [path].
Discussion
No questions or reviews yet.
Related
PluginsA skills library that gives coding agents a development process: brainstorming, planning, TDD, subagents and code review
Eleven skills take an agent from an idea to verified code through a PRD and RFCs kept as files in the repository
Two skills for architecture review and small safe refactorings: layer boundaries, SOLID and code smells in Kent Beck's manner.
A plugin with three skills for large development work: project analysis, phased tasks tracked as GitHub Issues, parallel agents and code review
An architecture coach skill: questions lead to ADRs, a diagram and a plan, and a repository is audited from its code
Eleven skills for Claude Code and Codex: test audit, git worktree cleanup, plan pages, screen recording transcripts and 1Password sign-in
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| PAAD (this card) | no | 153 | |
| Superpowers | no | 316,679 | a general set of agent development practices |
| Final Review | no | 113 | a single final check in place of a cycle of skills |
| Senior Software Architect Review | no | 251 | architecture review only, in one skill |
README in short
The README opens with an engineering-first thesis, then covers installation for Claude Code, the Claude desktop app and 70 other agents, tables of main and experimental skills and a configuration section. A full description of each skill lives in the tutorial on the author's site.