agentleFS
DnsQueenDnsQueen/orcinus-skills/skills/scaffold/SKILL.md

scaffold

Skill1 starsChanged 6 days agoMITView on GitHub

Set up a new project so build, lint, typecheck, and test all run from day one. Use when a calibrated idea is confirmed and there is no project yet, or when an existing project has no working checks. Only invoke via the orcinus router.

  • Reads credentials

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

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. 0 are common, 11 are distinctive, and 2 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.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    Create the smallest project that builds, lints, typechecks, and tests.

    line 10
  2. Distinctive:

    No scaffold before a calibrated brief.

    line 14
  3. Distinctive:

    Do not choose a stack silently.

    line 15
  4. Distinctive:

    Every file must earn its place.

    line 16
  5. Very specific:

    No real secrets, ever. Placeholders only, in .env.example.

    line 17
  6. Distinctive:

    Create only these, unless the stack needs one more file — and if so, name the need:

    line 28
  7. Distinctive:

    No test weakening. Never delete, weaken, or skip a test to make the suite pass.

    line 43
  8. Distinctive:

    No secrets in the repo.

    line 45
  9. Very specific:

    Put placeholder names in .env.example only.

    line 52
  10. Distinctive:

    Run, and record the real output and exit code for each:

    line 60
  11. Distinctive:

    All four must pass on the empty project with its one passing test.

    line 67
  12. Distinctive:

    Only now, commit with a clear message.

    line 71
  13. Distinctive:

    Not the feature. Not a stack choice made for the user.

    line 75

“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 — a project that is green from day one

Create the smallest project that builds, lints, typechecks, and tests. Do not write the real feature here. Do not commit a red project. The first commit happens only after all four commands pass.

Rules

  • No scaffold before a calibrated brief. If idea-calibrate has not confirmed the problem, user, success criteria, non-goals, and first slice, stop and run it first.
  • Do not choose a stack silently. Confirm it with the user in one round, with a recommended answer, and record the choice and why.
  • Every file must earn its place. No file without a stated need.
  • No real secrets, ever. Placeholders only, in .env.example. Never commit a secret, and never ask for one in chat.
  • The four commands are the gate. If any fails, fix only the cause and re-run all four.

Steps, in order

1. Confirm the brief and the stack

Read the calibrated brief. Confirm the first slice. Then ask the user to confirm the stack (language, test runner, linter). Record the answer. If the user has no preference, recommend TypeScript with tsc and Node's built-in test runner, and say why: no extra test framework to install, and the four commands stay simple.

2. Create the tree

Create only these, unless the stack needs one more file — and if so, name the need:

  • README.md — what the project is, in one paragraph, plus how to run the four commands.
  • package.json (or the stack's manifest) — with scripts named exactly build, lint, typecheck, test.
  • src/ — with one minimal module for the first slice (a stub that does nothing yet, but compiles).
  • tests/ — with one passing test for that stub.
  • docs/constitution.md — three to five invariants the project will enforce (see step 3).
  • .gitignore — node_modules/, dist/, .env, and the stack's build output.
  • .env.example — placeholder names only, no real values.
  • LICENSE — MIT unless the user says otherwise.

3. Write the constitution

In docs/constitution.md, write the invariants in plain language, for example:

  • No test weakening. Never delete, weaken, or skip a test to make the suite pass.
  • One slice at a time. One behaviour, one test, one commit.
  • No secrets in the repo.
  • Implementer is never the final judge; orcinus-verify checks the work.

These are rules, not wishes. Later work must be able to point at them.

4. Secrets and environment

Put placeholder names in .env.example only. Put .env in .gitignore and never commit it. If a service needs a real key, the user provides it on their own machine. Human-only steps (accounts, keys, provisioning) belong to the user, never to the agent.

5. Initialise git

git init -b main. Do not commit yet.

6. Verify — all four, in order

Run, and record the real output and exit code for each:

  1. build
  2. lint
  3. typecheck
  4. test

All four must pass on the empty project with its one passing test. If any fails, fix only the cause and re-run all four from the top. Do not skip a failing check and do not weaken the test to get a pass.

7. First commit

Only now, commit with a clear message. Record in the handoff: what was built, the four commands and their outputs, the stack choice and why, and what the first real slice will be.

What scaffold is not

Not the feature. Not a stack choice made for the user. Not a place to hide a failing check. If the four commands cannot all pass, the scaffold is not done.

Sources

Method, in our own words, inspired by: spec-kit (github/spec-kit, MIT) for the spec and constitution pattern; create-mercato-app (MIT) for the scaffolding pattern; repomix (MIT) for the codebase map, which comes after scaffolding, not during it. No text was copied from those projects.

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.