agentleFS
ziffanziffan/auto-continue-cli-agent/.claude/skills/session-start/SKILL.md

session-start

Skill1 starsChanged 4 months agoApache-2.0View on GitHub

Ritual buka sesi kerja — baca docs status/aturan wajib (DECISIONS → MAP → CONVENTIONS → GOTCHAS → CONTEXT → ISSUES; PROJECT/ARCHITECTURE = referensi on-demand), buktikan pemahaman eksplisit, lalu usulkan rencana sesi. Jalankan di AWAL setiap sesi pada proyek yang punya docs/, saat user bilang "mulai sesi", "lanjutkan proyek", "lanjut milestone", "kita di mana", atau saat sesi baru dibuka tanpa konteks pekerjaan sebelumnya.

39 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use deploy-verify, reflect and 2 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.
  • danielarif26Skill

    deploy-verify

    Verify a Vercel deployment actually went live and serves the correct production domain. Self-improving: reads its own run history before checking and logs the outcome after. Use when the user asks "did you deploy?", "is it live?", "verify the deploy", "check vercel", or after any Vercel deploy. Part of the Agentic OS (Level 1 backbone).

    Also in 3 projects that use this

  • Microsoft AzureSkill

    reflect

    Learning capture system that extracts HIGH/MED/LOW confidence patterns from conversations to prevent repeating mistakes. Use after user corrections ("no", "wrong"), praise ("perfect", "exactly"), or when discovering edge cases. Complements .squad/agents/{agent}/history.md and .squad/decisions.md.

    Also in 4 projects that use this

  • ibahgatSkill

    autopilot

    Run autonomous multi-stage execution loops until acceptance criteria pass or blockers are reached.

    Also in 3 projects that use this

  • MicrosoftSkill

    handoff

    Compact the current conversation into a handoff document for another agent to pick up.

    Also in 3 projects that use this

The file

Files
SKILL.md

Session Start — Handover In

Kenapa ritual ini ada: semua konteks proyek hidup di file, bukan di ingatan agent. Sesi yang langsung menyentuh kode tanpa membaca docs akan menciptakan konvensinya sendiri, me-relitigasi keputusan lama, dan mengulang gotcha yang sudah dibayar mahal.

Step 0 — Sinkron dengan remote (WAJIB, SEBELUM baca docs)

Proyek ini lintas-mesin (Ubuntu daily + Windows weekend) — mesin/sesi lain bisa sudah push ke origin. Memulai kerja di basis yang basi → commit di atas base lama → konflik push + kolisi nomor (ADR/G-/I-), rebase menyakitkan. Cegah di awal:

  1. git fetch origin (jangan asumsikan lokal terkini).
  2. Cek divergensi: git rev-list --left-right --count HEAD...origin/main (atau git status).
    • Lokal di belakang / diverge → integrasikan DULU (git pull --rebase bila lokal bersih; bila ada commit lokal, rebase ke origin/main) sebelum menyentuh tugas. Baca commit baru origin (mungkin ada audit/keputusan/renumber yang mengubah rencana).
    • Sinkron / lokal di depan bersih → lanjut.
  3. Bila working tree kotor saat fetch menunjukkan divergensi → laporkan ke user, jangan paksa.

Jangan lewati langkah ini walau "cuma mau baca" — status di CONTEXT.md lokal bisa ketinggalan dari kebenaran di origin.

Step 1 — Baca berurutan

Wajib baca penuh (status + aturan yang berubah tiap sesi — inti "di mana kita"):

  1. docs/DECISIONS.md — locked / pending / JANGAN PERNAH
  2. docs/MAP.md — peta repo
  3. docs/CONVENTIONS.md — pola wajib & terlarang
  4. docs/GOTCHAS.md — jebakan yang sudah ditemukan (indeks 1-baris)
  5. docs/CONTEXT.md — status terkini, milestone aktif, blocker (entri teratas = di mana kita)
  6. docs/ISSUES.md — issue open, prioritas (fokus bagian Terbuka)
  7. HANDOFF_CONTEXT_v{n}.md versi terbaru, bila ada

Referensi on-demand (spec/bentuk-sistem STABIL — jangan baca penuh tiap sesi; proyek matang tanpa milestone aktif per 19 Jul). Skim header untuk orientasi; deep-read HANYA saat task hari ini menyentuhnya (fitur baru, ubah arsitektur, sengketa scope):

  • docs/PROJECT.md — masalah, persona, user story, AC. Flow + wireframe sudah diarsip (.archived/PROJECT-design.md) — baca dari sana bila perlu.
  • docs/ARCHITECTURE.md — C4, container map, tech stack.

Baca untuk dipakai, bukan untuk diringkas panjang — jangan paste ulang isinya ke chat.

Step 2 — Bukti pemahaman (wajib SEBELUM menyentuh tugas)

Setor ke user, ringkas:

  1. Status proyek dalam maksimal 3 kalimat.
  2. Stack + versi yang di-lock.
  3. Dua gotcha paling relevan untuk pekerjaan hari ini.
  4. Satu locked decision yang paling membatasi pekerjaan hari ini.
  5. Blocker aktif + pending decision yang lewat deadline (bila ada — tagih ke owner-nya).

Tujuannya memaksa read-before-act. Kalau tidak bisa mengisi kelima poin, berarti bacaan Step 1 belum cukup — ulangi, jangan menebak.

Step 3 — Usulkan rencana sesi

  • Ambil dari milestone aktif di CONTEXT.md + slice berikutnya di MILESTONES.md.
  • Usulkan 1-2 sub-task (1-3 jam kerja per sub-task), sebutkan bukti verifikasi yang akan disetor di akhir tiap sub-task.
  • Satu sesi = satu modul. Jangan multitask antar modul — context switch mahal untuk user dan untuk agent. Budget konteks sesi: ±300K token soft; rencanakan agar muat.
  • Tunggu approval user sebelum mulai eksekusi.

Kondisi khusus

  • docs/ tidak ada atau kosong → jangan lanjut; tawarkan skill docs-first-spec.
  • CONTEXT.md basi (tidak cocok dengan git log terakhir) → laporkan selisihnya, rekonstruksi status dari git log + ISSUES.md, lalu minta konfirmasi user.
  • Spec belum LOCKED tapi user minta implementasi → ingatkan, tawarkan menyelesaikan lock dulu via docs-first-spec Step 7.

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.