본문으로 건너뛰기
cccskills
무료GitHub에 공개

ship

Lint, build, test, verify specs, commit, and push — one micro-commit to remote. Validates the working tree, verifies spec alignment, and creates an atomic save point.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

설치 방법 보기

포함된 파일(1)

  • SKILL.md4.6 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

Ship — validate, commit, and push micro-commits

Validate the working tree, verify spec alignment, create a micro-commit, and push it to the remote — a single atomic "save point."

Trigger

When a small, bounded unit of work is completed and needs to be shipped safely.

Goal

Ensure the codebase is green (lint, build, test, typecheck), verify that changed code stays aligned with its specs, create a well-formed micro-commit, and push to the current remote branch.

Boundaries

  • In scope: running quality gates, verifying spec compliance, updating assessment retrospective, creating and pushing atomic commits.
  • Out of scope: fixing lint/type/test failures (dev), authoring or amending specs (spec), creating Linear issues (lead).

Pipeline

Execute these steps in order. Stop immediately on any failure — do not skip steps or push broken code.

Step 1 — Inventory

Run in parallel:

  • git status — identify changed files
  • git diff --staged and git diff — understand the actual changes
  • git log --oneline -5 — read recent commit style

If there are no changes (clean working tree), stop and tell the user.

Step 2 — Quality Gates

Run in parallel:

  • pnpm lint
  • pnpm build
  • pnpm test:run
  • pnpm check

All four must pass. On failure: report which gate failed with the error output and stop. Do not attempt to fix — that is the developer's job.

Step 3 — Spec Verification

For every changed file, check whether a related spec exists under specs/. Use the file path and component name to locate relevant specs.

For each relevant spec found:

  1. Read the spec's Regression Guardrails and Definition of Done sections.
  2. Verify that the changes do not violate any guardrail.
  3. Verify that the changes do not regress any Definition of Done item that was previously satisfied.
  4. If the change modifies behavior covered by a spec, verify the spec was updated in the same changeset (same-commit rule).

If a spec violation or missing spec update is found: report it and stop. Do not commit.

If no spec exists for the changed files: that's fine — not all files have specs. Note it and proceed.

Step 3.5 — Assessor Retrospective (Double-Loop Learning)

For all changed files identified in Step 1, check if there are any related assessments in docs/assessments/ledger.jsonl with execution_status: "pending". An assessment is related if its slug, target files, or challenger title overlap with the files being committed.

If matching pending assessments are found:

  1. Prompt for Retro Feedback: Ask the user if the implementation of the assessment succeeded, and ask for any unforeseen difficulties or lessons.
  2. Update the Ledger: Update the corresponding JSON line in docs/assessments/ledger.jsonl:
    • "execution_status": "success" (or "failed")
    • "execution_retro": A brief summary of the implementation results (e.g. build results, user feedback)
    • "lessons_learned": Any specific lessons or guidelines surfaced.
  3. Consolidate Lessons: Update docs/assessments/lessons.md with the new heuristics under ## Heuristics & Lessons.
  4. Stage Retro Files: Stage the updated docs/assessments/ledger.jsonl and docs/assessments/lessons.md so they are automatically included in the commit.

Step 4 — Stage & Commit

  1. Stage all relevant changed files with git add (specific files, not -A).
  2. Draft a commit message:
    • Use conventional commit format: type(scope): description
    • Keep the subject line under 72 characters
    • Match the style of recent commits from Step 1
    • If the user provided a message via $ARGUMENTS, use it as the basis
  3. Create the commit.

Step 5 — Push

  1. Check if the current branch tracks a remote. If not, push with -u origin HEAD.
  2. Push to the remote.
  3. Report success: branch name, commit hash (short), and commit message.

Principles

  • Atomic save points — Each /ship is one micro-commit. Small, reversible, traceable.
  • Green before push — Never push code that fails any quality gate.
  • Specs are law — Changed behavior must have a matching spec update. No silent drift.
  • No fixes in flight — If something fails, stop and report. The developer fixes; then runs /ship again.
  • No force push — Always a regular push. If the remote has diverged, stop and tell the user.

Instructions

$ARGUMENTS

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

assemble

무료

Dev-Critic loop: implement, review, fix — repeat until clean

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

villetakanen/asdlc-io452026년 10월 9일 업데이트

assess

무료

Assess source material against the ASDLC Knowledge Base and produce an evidence-calibrated Content Review Report.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

villetakanen/asdlc-io452026년 10월 9일 업데이트

critic

무료

Perform an adversarial review of the current uncommitted changeset, validating it against project contracts and architectural intent.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

villetakanen/asdlc-io452026년 10월 9일 업데이트

curator

무료

Triage the content corpus against the latest GSC snapshot and produce a prioritized refresh report. Read-only: reports, never edits.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

villetakanen/asdlc-io452026년 10월 9일 업데이트

dev

무료

Implement features, fix bugs, and keep the codebase healthy and maintainable. Works from a defined PBI, issue, or task description.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

villetakanen/asdlc-io452026년 10월 9일 업데이트

geo-audit

무료

GEO+SEO audit of asdlc.io — analyzes the codebase for Generative Engine Optimization across AI citability, technical infrastructure, content quality, schema markup, and knowledge base discoverability. Produces a composite GEO Score (0-100) with prioritized action plan.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

villetakanen/asdlc-io452026년 10월 9일 업데이트

villetakanen의 스킬 모두 보기

이 스킬의 문제 신고