gwpicard/ai-build-kit
A workflow for building and running internal tools with a coding agent.
field: unknown workflow: PR / code reviewGit & commitsTesting & TDDCI/CD & deployResearch runs claude in CI
ARCHETYPE
Skill Collector
Ten-plus skills loaded on demand. A procedural-knowledge library.
Copy this rig
# review before running: this installs third-party code
$ npx degit gwpicard/ai-build-kit/.claude ./rig-ai-build-kit # inspect, then merge into .claude/ MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ npx degit gwpicard/ai-build-kit/.agents/skills/build-piece .claude/skills/build-piece $ npx degit gwpicard/ai-build-kit/.agents/skills/clarify .claude/skills/clarify $ npx degit gwpicard/ai-build-kit/.agents/skills/evidence-run .claude/skills/evidence-run $ npx degit gwpicard/ai-build-kit/.agents/skills/fit-check .claude/skills/fit-check $ npx degit gwpicard/ai-build-kit/.agents/skills/implement .claude/skills/implement $ npx degit gwpicard/ai-build-kit/.agents/skills/live-copy .claude/skills/live-copy $ npx degit gwpicard/ai-build-kit/.agents/skills/maintain .claude/skills/maintain $ npx degit gwpicard/ai-build-kit/.agents/skills/merge .claude/skills/merge $ npx degit gwpicard/ai-build-kit/.agents/skills/quality-checks .claude/skills/quality-checks $ npx degit gwpicard/ai-build-kit/.agents/skills/records .claude/skills/records $ npx degit gwpicard/ai-build-kit/.agents/skills/repair .claude/skills/repair $ npx degit gwpicard/ai-build-kit/.agents/skills/research .claude/skills/research $ npx degit gwpicard/ai-build-kit/.agents/skills/screen-check .claude/skills/screen-check $ npx degit gwpicard/ai-build-kit/.agents/skills/second-opinion .claude/skills/second-opinion $ npx degit gwpicard/ai-build-kit/.agents/skills/setup-ai-build-kit .claude/skills/setup-ai-build-kit $ npx degit gwpicard/ai-build-kit/.agents/skills/setup-hosting .claude/skills/setup-hosting $ npx degit gwpicard/ai-build-kit/.agents/skills/shape .claude/skills/shape $ npx degit gwpicard/ai-build-kit/.agents/skills/triage .claude/skills/triage $ npx degit gwpicard/ai-build-kit/.agents/skills/what-now .claude/skills/what-now
$ curl -fsSL --create-dirs -o .claude/agents/builder.md https://raw.githubusercontent.com/gwpicard/ai-build-kit/stable/.claude/agents/builder.md $ curl -fsSL --create-dirs -o .claude/agents/refuter.md https://raw.githubusercontent.com/gwpicard/ai-build-kit/stable/.claude/agents/refuter.md $ curl -fsSL --create-dirs -o .claude/agents/researcher.md https://raw.githubusercontent.com/gwpicard/ai-build-kit/stable/.claude/agents/researcher.md $ curl -fsSL --create-dirs -o .claude/agents/reviewer.md https://raw.githubusercontent.com/gwpicard/ai-build-kit/stable/.claude/agents/reviewer.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"permissions": {
"deny": [
"Bash(git reset --hard:*)",
"Bash(git push --force:*)",
"Bash(git push -f:*)",
"Bash(git clean -f:*)",
"Bash(git clean -fd:*)",
"Bash(rm -rf:*)"
]
}
} Skills (19)
build-piececlarifyevidence-runfit-checkimplementlive-copymaintainmergequality-checksrecordsrepairresearchscreen-checksecond-opinionsetup-ai-build-kitsetup-hostingshapetriagewhat-now
Subagents (4)
| builder model: inherit | Builds one ready piece in a fresh context, for an auto run of several pieces. Runs the build-piece skill on the branch it is given, never merges, and reports back. Started by implement, never by the p |
| refuter model: inherit | A fresh reviewer who tests whether one review finding is wrong. Used only for findings marked worth stopping for. |
| researcher model: inherit | Settles one outside fact against its primary source, or searches for existing work, in a fresh context, and returns a short cited note. Runs the research skill. Started by a command, never by the pers |
| reviewer model: inherit | Fresh eyes on the code and evidence for a piece on the pull-request route. Runs the second-opinion skill in a context that has not seen the builder's reasoning. |
Slash commands (6)
/implement/maintain/setup-ai-build-kit/setup-hosting/shape/what-now
Permissions
deny (6)
Bash(git reset --hard:*)
Bash(git push --force:*)
Bash(git push -f:*)
Bash(git clean -f:*)
Bash(git clean -fd:*)
Bash(rm -rf:*)
ask (0)
—
allow (0)
—
Similar rigs
Freakling/App-Director
Director-oriented AI coding workflow for small apps. You direct the design, AI builds it. Works for web, mobile, and desktop apps.
Fort Knox 2.3k tok · · Q41
Freakling/Director-Driven-Development
How-to docs for Director-Driven Development: you direct, AI agents build, one automatic check proves every task. Covers Godot-Director, App-Director and SaaSAllTheThings. Built with App-Director.
Fort Knox 4.1k tok · · Q56
iuliandita/agents
Global prompts and subagent roles for coding agents, deployed to every harness you use.
Orchestrator 1.5k tok · · Q3
Freakling/Godot-Director
Keep designing your Godot 4 game while AI builds it. Design calls come to you as options, one check proves every change works, and your game's git repo holds all the state. Built for Claude Code; any assistant that reads AGENTS.md can use i
Fort Knox 4.0k tok · · Q52