Skip to content
Back to skills

Setup Ai Build Kit

ASecurity

Start or resume project setup with /setup-ai-build-kit. Use /implement after founding, or /shape for features and repairs.

  • 2 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 21, 2026
developmentpythonapidatabase

Works with

  • api

Security analysis

A100/100

Pro scans all 20 files and shows the line behind each finding

Scanned October 21, 2026

npx -y skills add gwpicard/ai-build-kit --skill setup-ai-build-kit --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Setup Ai Build Kit?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Setup Ai Build Kit
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/gwpicard-setup-ai-build-kit/badge)](https://www.skillsdirectory.com/skills/gwpicard-setup-ai-build-kit)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: setup-ai-build-kit
description: Start or resume project setup with /setup-ai-build-kit. Use /implement after founding, or /shape for features and repairs.
---

# Setup AI Build Kit

Prepare a project without writing feature code.

## Conversation contract

Speak when the person must answer, decide, approve, understand a blocker, or
see a result that changes what they can do next. Finishing an internal step is
not a user-facing result by itself. Routine reading, research, setup, and
checking happen quietly. Ask every question in the shape clarify's rules
describe. Name a service only when the person must choose, own its account,
grant access, pay, or understand a limit; say what it is in one plain sentence
at the same moment.

## 1. Orientation

Before technical work, explain the private setup and checkpoint, and that
nothing will be published without approval. For existing work, say you will
first understand it. Explain permissions just before a box. Do not announce
that a confirmation might appear later.

Founding ends with a saved project. Record open questions with your best guess
and check gaps as tasks. If they say to continue, state assumptions and carry
on.

## Route

1. Load `references/resume.md`.
2. Run `scripts/check-tooling.sh` and follow `references/capability-check.md`.
   Ask whether to adopt existing work; if yes, use `references/adopting.md`.
3. Follow `references/need-for-software.md`, then run the `clarify` skill. Settle terms,
   facts and visual questions with evidence or a prototype.
   Keep questions visible. Save answers in ignored `.agents/tmp/setup-notes.md`
   before the next question. Load `references/stand-up.md` before the closing
   reply and include its menu.
   Python document-worker and existing-database recipe parts are not offered
   yet; `Fits:` lists tool shapes.
4. Run `fit-check`: its three questions, then access, risk areas, cautions
   and recheck triggers. Consider redesigns first. Give the risk notice before
   flagged work and record acceptance. An ownership no becomes a task.
5. Follow `references/write-the-masterplan.md`. If a risk area is named, run
   `second-opinion` before flagged work. Record findings and gaps; apply a
   stopping finding only with approval. Otherwise record why review was
   skipped. Check ownership.
6. Cut one-sitting vertical slices in dependency order, not database, API and
   UI pieces. Follow `references/cut-the-plan.md` to save them as issues.
7. Follow `references/stand-up.md` to prepare, check and save; report with
   `references/completion-report.md`.

Do not push or open a pull request during founding. The first upload waits for
the approval required by `merge`.

## Done

Finish with `scripts/check-tooling.sh --conditions` at exit 0, records and
review, visible pieces, a passing check and a saved local checkpoint.
Record private Free limits; founding continues. End with
`/implement` and `/shape` in a fresh chat.

Files in this skill

  • SKILL.md2.9 KB
  • references/adopting.md2.1 KB
  • references/blocked-commands.md1.8 KB
  • references/capability-check.md3 KB
  • references/codex-github.md3.8 KB
  • references/completion-report.md2.9 KB
  • references/cut-the-plan.md3.6 KB
  • references/manual-setup.md1.5 KB
  • references/need-for-software.md805 B
  • references/required-tools.md2.2 KB
  • references/resume.md2.1 KB
  • references/stand-up.md3.8 KB
  • references/write-the-masterplan.md1.4 KB
  • scripts/bootstrap-project.sh9.2 KB
  • scripts/check-tooling.sh16 KB
  • scripts/codex-with-github.py2.3 KB
  • scripts/found-issues.sh7.1 KB
  • scripts/found-save.sh7 KB
  • scripts/place-plan-helper.sh3.8 KB
  • scripts/place-trim-check.py8.2 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.