agentleFS
awahidsawahids/noskill/skills/project-profile/SKILL.md

project-profile

Skill1 starsChanged 26 days agoMITView on GitHub

Scans the repo once and writes project-profile.md (stack, test/lint/typecheck commands, folders, helpers, rules) so agents don't re-explore per ticket. Use on new projects, when feature-flow finds no profile, or for "profil project", "scan project".

6 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use tdd.

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.
  • GoogleSkill

    tdd

    Test-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 1 is distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read only manifest and config files, not all source: package.json, composer.json, go.mod, pyproject.toml, Makefile, docker-compose*, CI …

    line 15
  2. Very specific:

    Verify commands: run the test command (and typecheck/lint if present) once with limited output (| tail -n 5).

    line 18
  3. Very specific:

    If the fast version fails or the tool version doesn't support it, write belum dicek and use the normal version.

    line 25
  4. Very specific:

    Only facts verified from files or command output.

    line 77
  5. Distinctive:

    If the repo is git, suggest the user commit this file so the next session uses it directly.

    line 83

“Common” is measured: kept by at least 6 in 10 of the 5 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

Project Profile

noskill docs folder: ${user_config.docs_dir} — written as <docs> below. If that value still appears literally (not substituted), use docs/noskill.

Goal: one compact file every agent reads before working, so test commands, folder patterns, and helpers aren't searched again on every ticket. Use when starting work on a new project, when feature-flow finds no profile, or when the user asks.

How to scan (token-efficient)

  1. Read only manifest and config files, not all source: package.json, composer.json, go.mod, pyproject.toml, Makefile, docker-compose*, CI files (.github/workflows/*), CLAUDE.md, the top of README.
  2. Folder structure: ls the top two levels (ignore node_modules, vendor, .git, build folders).
  3. Common helpers: grep exported function names in utility folders (lib/, utils/, helpers/, pkg/, internal/*/util*, app/Helpers). Record the name and a one-line purpose; don't read them in full.
  4. Verify commands: run the test command (and typecheck/lint if present) once with limited output (| tail -n 5). Record whether it passes, how long it takes, and how to run a single test file.
  5. Fast versions for per-ticket use, with cache in .noskill-cache/ (add .noskill-cache/ to .git/info/exclude, not .gitignore, so it isn't committed). Check once that the fast version works, then record:
    • TypeScript: npx tsc --noEmit --incremental --tsBuildInfoFile .noskill-cache/tsc.tsbuildinfo (second and later runs only recheck changed files)
    • ESLint: npx eslint --cache --cache-location .noskill-cache/eslint <file>
    • Jest: npx jest --findRelatedTests <file>; Vitest: npx vitest related <file> --run
    • Go: go vet ./<paket>/... and go test ./<paket>/... (Go already uses its own cache)
    • PHP: PHPStan and Pest/PHPUnit with a file filter or related tests If the fast version fails or the tool version doesn't support it, write belum dicek and use the normal version.

<docs>/project-profile.md format (max ±80 lines)

# Project Profile — <nama>
_Dibuat: <tanggal> · sumber: <commit hash pendek>_

## Stack
- Bahasa/framework/versi, database, ORM, package manager

## Command
| Tujuan | Command | Status saat scan |
|---|---|---|
| Install | `npm ci` | — |
| Test semua | `npm test` | lulus, 77 test, ±5 dtk |
| Test satu file | `npx tsx --test lib/goals.test.ts` | — |
| Test terkait file yang diubah | `npx jest --findRelatedTests <file>` | — |
| Test tanpa DB (per tiket) | `npx jest --testPathIgnorePatterns integration` | lulus, tanpa DB/server |
| Test integrasi (butuh DB, Stage 6) | `npx jest integration` | belum dicek |
| Lint file tertentu | `npx eslint --cache --cache-location .noskill-cache/eslint <file>` | — |
| Jalankan environment (DB + server, Stage 6) | `docker compose up -d db && npm run start:dev` | — |
| Health check | `curl -sf localhost:3000/health` | — |
| Typecheck | `npx tsc --noEmit` | lulus |
| Typecheck cepat (per tiket) | `npx tsc --noEmit --incremental --tsBuildInfoFile .noskill-cache/tsc.tsbuildinfo` | lulus, ±3 dtk setelah run pertama |
| Lint | `npm run lint` | lulus |
| Folder dependency (tidak di-commit) | `node_modules/` | — |

## Struktur
- `app/` — halaman Next.js (App Router)
- `lib/` — logic bisnis + unit test (`*.test.ts` di sebelah file)

## Helper yang dipakai ulang
- `lib/currency.ts` — `formatMoney(n)`: format Rupiah
- `lib/date.ts` — `todayISO()`: tanggal hari ini YYYY-MM-DD

## Aturan project
- (dikutip singkat dari CLAUDE.md/README/CI, misal: setiap perubahan yang terlihat user menambah item di `lib/changelog.ts`)

## Repo (hanya kalau fitur menyentuh beberapa repo)
| Nama | Path | Profil |
|---|---|---|
| api | . | project-profile.md |
| web | ../web-client | project-profile-web.md |

## Integrasi
- tracker: none            # none | github | notion — tujuan sinkron board tiket
- notion_database: —       # URL/ID database Notion kalau tracker: notion

Rules

  • Only facts verified from files or command output. Anything uncertain is written belum dicek, never guessed.
  • Refresh if a manifest/lockfile changed since the commit in the header, if a command in the profile fails, or if the user asks. Replace only the changed parts, and add rows from the current format that the profile lacks (e.g. Jalankan environment, Health check, fast per-ticket commands) — an old profile may predate them. Keep everything else, including lines the user wrote by hand.
  • DB-free vs integration tests: find how the project separates tests that need a database (folder, file suffix, tag, or Jest/Vitest project) and record one command for each. If it doesn't separate them, write belum dipisah — per-ticket runs then use Test terkait file yang diubah and treat specs that fail for lack of a database as Stage 6 specs.
  • Environment row: find the command in docker-compose*, package scripts, Makefile, or README. Verify it once if it can run here (start, Health check succeeds, stop); otherwise write the command with belum dicek.
  • The tracker field is filled from the user's answer (default none). Don't ask again once it's filled.
  • Other repos (multi-repo): scan that repo the same way (path from the Repo section), but save the result in the docs repo as <docs>/project-profile-<nama>.md without the Repo and Integrasi sections. Other repos get no noskill folder.
  • If the repo is git, suggest the user commit this file so the next session uses it directly.

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.