agentleFS
theyoungastronautstheyoungastronauts/polaris/skills/scaffold/SKILL.md

scaffold

Skill4 starsChanged 3 months agoMITView on GitHub

Scaffold sub-project directories from an approved design/plan — git init and run the stack bootstraps. Use when turning a plan into real project structure.

  • Installs packages

49 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use ascii-diagram, docs-ci and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • SethGammonSkill

    ascii-diagram

    Generate perfectly aligned ASCII diagrams — architecture, flow, sequence, box-and-arrow. Uses a programmatic character-grid approach so alignment is guaranteed by math, not token prediction. Includes post-render verification.

    Also in 4 projects that use this

  • jh941213Skill

    docs-ci

    문서 CI(링크·OpenAPI lint·breaking 게이트·mermaid 검증)나 문서 드리프트 감지(docs.yaml 신선도 검사)를 설치하거나 돌릴 때 사용. 문서 내용 작성은 docs-architecture·docs-interfaces·docs-manuals.

    Also in 3 projects that use this

  • jh941213Skill

    docs-interfaces

    API·이벤트 인터페이스 문서(OpenAPI·swagger·AsyncAPI 명세, 엔드포인트 레퍼런스, API 구성도·시퀀스, API CHANGELOG)를 만들거나 코드와 맞출 때 사용. 시스템 구성도는 docs-architecture, API 설계 원칙은 api-design-principles.

    Also in 3 projects that use this

  • SethGammonSkill

    research-fleet

    Parallel research using Fleet wave mechanics. Spawns multiple scout agents, each investigating a different angle of the same question. Findings are compressed between waves. Produces a unified research brief from multiple independent perspectives.

    Also in 3 projects that use this

  • RachaelQuiselSkill

    client-docs

    Client documentation. Use when a build needs documentation updates or missing project documents need to be identified.

    Also in 3 projects that use this

  • vicoa-aiSkill

    live-preview

    Start a local app in the current user project, expose it through a tunnel, and return preview details in structured Markdown. Use when the user asks for a preview URL, public URL, shareable localhost, mobile/web preview, or tunnel setup.

    Also in 4 projects that use this

  • +4 more from jh941213

The file

We split this skill into the rules it gives an agent

11 rules. 0 are common, 11 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 42 edited versions we read keep these.

  • 11

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After brainstorming is complete and you have a design doc.

    line 11
  2. Distinctive:

    Read CLAUDE.md for the selected stacks and the design doc for what's being built.

    line 24
  3. Distinctive:

    Before creating anything, collect all bootstrap inputs from the user so teammate agents can run non-interactively.

    line 56
  4. Distinctive:

    Use the design doc and project name to propose sensible defaults for most of these.

    line 82
  5. Distinctive:

    When there are multiple sub-projects, bootstrap them in parallel using a team.

    line 123
  6. Distinctive:

    All configuration was gathered in step 3 — agents receive pre-filled values and should not prompt the user for any bootstrap inputs.

    line 126
  7. Distinctive:

    Wait for all agents to complete

    line 144
  8. Distinctive:

    Use the pre-filled configuration values from step 3

    line 151
  9. Distinctive:

    After bootstrap completes, make an initial commit:

    line 152
  10. Distinctive:

    After bootstrapping, check if Axon is installed and index the project for structural code intelligence.

    line 173
  11. Distinctive:

    Do not block scaffolding on Axon — it's a recommendation, not a requirement.

    line 197

“Common” is measured: kept by at least 6 in 10 of the 42 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Scaffold: Design to Projects

When to Use

After brainstorming is complete and you have a design doc. Scaffold creates the project structure so the planner has real directories, installed packages, and boilerplate to reference when building the implementation plan.

Prerequisites

  • A brainstorm/design doc exists (docs/plans/*.md)
  • Stack context is already in CLAUDE.md (set up by polaris new)

Process

1. Identify Sub-Projects from the Design

Read CLAUDE.md for the selected stacks and the design doc for what's being built. Map each sub-project to a bootstrap command:

Sub-project typeStack flagBootstrap command
Django/DRF API--stack django/django-bootstrap
Next.js frontend--stack nextjs/nextjs-bootstrap
Flutter app--stack flutter(manual setup)

2. Confirm with the User

Present the scaffold plan before creating anything:

Scaffold Plan:
  Root: ~/prj/my-app/
  Stacks (from CLAUDE.md):
    - Backend: Django → server/
    - Frontend: Next.js → web/

  Will create:
    - Subdirectories within the current project
    - git init each sub-project
    - Run bootstrap commands

  Parallel bootstrap: Yes (2 sub-projects, no shared code)

Let the user adjust names, add/remove stacks, or change directories before proceeding.

3. Gather Bootstrap Configuration

Before creating anything, collect all bootstrap inputs from the user so teammate agents can run non-interactively. Bootstrap commands ask interactive questions — if multiple agents prompt simultaneously, it's unusable.

For each sub-project, ask the user for the required values upfront:

Django bootstrap values:

  • Blueprint: minimal (Django + DRF + Postgres + Docker — no auth, no Redis), standard (+ Redis, custom User/JWT auth, Sentry — most production apps), or full (+ Celery/Beat, S3/R2, email, discord, Heroku)
  • Service/directory name (e.g., my_service)
  • Docker Compose project name (kebab-case, e.g., my-service)
  • Database name (snake_case, e.g., my_service)
  • Host port (default: 8000)
  • Cache key prefix (e.g., myserv)
  • Heroku app name (only if blueprint is full and deploying, otherwise skip)

Next.js bootstrap values:

  • App name (e.g., my-app)
  • Docker Compose project name (kebab-case, e.g., my-app-web)
  • Host port (default: 3000)
  • Backend API port (default: 8000 — match the Django port above)
  • Production domain (e.g., myapp.com)
  • App display title (e.g., My App)
  • Architecture mode: frontend-centric, SSR-centric, or combination

Note: When the Django blueprint is minimal (no auth), recommend the Next.js app without auth pages (no login, register, or protected route patterns). Auth can be added later if the Django project upgrades to standard.

Use the design doc and project name to propose sensible defaults for most of these. Present them as a table the user can confirm or override:

Bootstrap Configuration:

  Django (server/):
    Blueprint:       standard        (most production apps with auth)
    Service name:    my_service      (from project name)
    Compose project: my-service
    Database:        my_service
    Host port:       8000
    Cache prefix:    myserv
    Heroku app:      (skip — full blueprint only)

  Next.js (web/):
    App name:        my-app
    Compose project: my-app-web
    Host port:       3000
    API port:        8000            (matches Django above)
    Domain:          myapp.com
    Display title:   My App
    Architecture:    SSR-centric     (recommended for new projects)

  Look right? [Y/n]

4. Create and Initialize

For each sub-project:

mkdir -p {root}/{suffix}
cd {root}/{suffix}
git init

Naming convention: {suffix} = role (api, web, mobile, admin) as a subdirectory of the root.

5. Run Bootstrap Commands

When there are multiple sub-projects, bootstrap them in parallel using a team. Sub-projects have no shared code at this stage, so there are no conflicts.

All configuration was gathered in step 3 — agents receive pre-filled values and should not prompt the user for any bootstrap inputs.

Parallel bootstrap (2+ sub-projects):

  1. Spawn one named agent per sub-project using the Agent tool (called Task in older Claude Code versions). Give each a name: so you can address it with SendMessage:

    For each sub-project, spawn a general-purpose agent with the pre-filled config:

    You are bootstrapping the {label} sub-project at {root}/{suffix}/. Read the bootstrap skill at .claude/skills/{bootstrap_command}/SKILL.md and follow it. Use these pre-filled configuration values (do NOT prompt the user for these):

    blueprint={blueprint} {list all remaining key=value pairs from step 3 for this sub-project}

    After bootstrap completes, run: git add . && git commit -m "chore: initial {label} scaffold" Report back what was created.

  2. Wait for all agents to complete
  3. Send each agent a final message once all bootstraps are done

Single sub-project (no team needed):

  1. Change into the sub-project directory
  2. Invoke the bootstrap command (/django-bootstrap or /nextjs-bootstrap)
  3. Use the pre-filled configuration values from step 3
  4. After bootstrap completes, make an initial commit:
    git add . && git commit -m "chore: initial project scaffold"

6. Generate VS Code Workspace (Optional)

{
  "folders": [
    { "path": ".", "name": "Planning" },
    { "path": "api", "name": "API" },
    { "path": "web", "name": "Web" }
  ]
}

Save as {root}/{root-name}.code-workspace.

7. Index with Axon (if available)

After bootstrapping, check if Axon is installed and index the project for structural code intelligence.

# Check if axon is available
if command -v axon &>/dev/null; then
    # Index each sub-project
    axon analyze {root}/{suffix}
    # Repeat for each sub-project
fi

If Axon is not installed, warn the user:

Note: Axon (code intelligence) is not installed.
Axon provides structural analysis (call graphs, impact analysis, dead code detection)
that significantly improves planning and verification quality.

Install with: pip install axoniq   (or: uv add axoniq)
Then run: axon analyze .

See the axon-code-intel skill for details on how agents use it.

Do not block scaffolding on Axon — it's a recommendation, not a requirement.

8. Report Summary

Scaffold complete:
  ~/prj/my-app/              (project root)
  ~/prj/my-app/server/       (Django backend) -- ready
  ~/prj/my-app/web/          (Next.js frontend) -- ready
  ~/prj/my-app/my-app.code-workspace

Then suggest the next step:

Project is scaffolded. Next:

1. Start a NEW Claude session in this directory
2. Tell it: "Turn the design into a phased implementation plan"
   - The planner can now reference real project files and structure
3. Review the plan, then execute phase by phase

Key Principles

  • Confirm before creating -- always show the plan and get user approval first
  • Monorepo with subdirectories -- stacks live as subdirectories in a single repo
  • Scaffold before planning -- real project structure makes plans more concrete
  • Parallel when possible -- bootstrap sub-projects concurrently with named agents
  • Bootstrap commands do the heavy lifting -- this skill orchestrates; the bootstrap commands handle details

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.