project-bootstrap

name	project-bootstrap
description	Bootstrap a new project with Cursor-ready configuration: AGENTS.md, per-project .cursor/rules/project-conventions.mdc, and optional memory_bank/ scaffold. Use at project initialization, when starting a new repo, or when the user says "set up Cursor for this project" / "bootstrap this project" / "init project config".

When to use

• Starting a new project repo that needs Cursor-ready config
• User says "set up Cursor for this project", "bootstrap this project", or "init project config"
• A repo lacks AGENTS.md, .cursor/rules/project-conventions.mdc, or memory_bank/

On-demand loading: Read when initializing a new repo with Cursor config (AGENTS.md, .cursor/rules, memory_bank). Skip for changes inside ~/.cursor global setup.

project-bootstrap

Purpose

Generate the minimal Cursor configuration for a new or unconfigured project so that every agent and rule in ~/.cursor works correctly from the first prompt. Output is always tailored to the actual repo — never generic boilerplate.

---

When to activate

• /project-init command is run
• User says: "bootstrap this project", "set up Cursor for this repo", "init project config"
• SCOUT discovers no .cursor/rules/, no AGENTS.md, and no CLAUDE.md in the project
• Starting a new project from scratch

---

Step 1 — Detect project facts

Before writing anything, collect:

stack:          [e.g. Next.js 15 / TypeScript / Prisma / pnpm]
package_manager:[npm | pnpm | yarn | bun]
test_cmd:       [from package.json scripts or Makefile]
lint_cmd:       [from package.json scripts or eslint.config.*]
build_cmd:      [from package.json scripts]
framework:      [Next.js | React | Vue | Express | NestJS | Rails | Django | ...]
db:             [PostgreSQL | MySQL | MongoDB | SQLite | none]
git_branch:     [main | master | other]
existing_rules: [.cursor/rules/ present? yes/no]
existing_agents:[AGENTS.md or CLAUDE.md present? yes/no]

Use git, package.json, Makefile, Cargo.toml, go.mod, pyproject.toml, Gemfile, etc. to fill these. Do not guess — leave unknown if unconfirmed.

---

Step 2 — Generate AGENTS.md (at project root)

Create or update AGENTS.md with imperative learnings for this project. Max 30 items, one line each. Keep it dense and actionable. → Template with all sections in references/agents-md-template.md.

---

Step 3 — Generate .cursor/rules/project-conventions.mdc

Create .cursor/rules/project-conventions.mdc scoped to this project with stack facts, module boundaries, active patterns, anti-patterns, and validation commands. → Template in references/conventions-template.md.

---

Step 4 — Optionally scaffold memory_bank/

If the user asked to "set up memory" or if this is a significant project (evidence: existing docs/, multiple contributors, >10 source files), invoke the memory-bank skill to scaffold memory_bank/ in the project root.

---

Step 5 — Create .cursor/plans/ directory

Ensure .cursor/plans/ exists so SDLC artifacts can be written immediately.

mkdir -p .cursor/plans

---

Output summary

After completing, report:

Created:
  - AGENTS.md (project root)
  - .cursor/rules/project-conventions.mdc
  - .cursor/plans/ (directory)
  - memory_bank/ (if activated)

Detected facts:
  stack: [...]
  commands: install=[...] test=[...] lint=[...] build=[...]

Next: Run /ks-conductor for your first task — it will read these files automatically.

---

Maintenance guidance

• Keep AGENTS.md imperative and dense; prune stale learnings regularly
• Update project-conventions.mdc when architecture decisions change
• Run /retro-to-rule to convert retrospective findings into new convention entries
• Use /knowledge-consolidate to promote user-cursor10x-mcp memory into these files