yktsnet/sdlc-kit
Conventions for developing with AI agents, installed into your repo one unit at a time: workflow, guards, doc style.
stack: Shell
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
Copy this rig
# review before running: this installs third-party code
$ npx degit yktsnet/sdlc-kit/units/customer/.claude ./rig-sdlc-kit # inspect, then merge into .claude/ MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ npx degit yktsnet/sdlc-kit/units/customer/.claude/skills/qa-collect .claude/skills/qa-collect $ npx degit yktsnet/sdlc-kit/units/customer/.claude/skills/qa-send .claude/skills/qa-send $ npx degit yktsnet/sdlc-kit/units/daily-report/.claude/skills/daily-report .claude/skills/daily-report $ npx degit yktsnet/sdlc-kit/units/doc-style/.claude/skills/jp-writing .claude/skills/jp-writing $ npx degit yktsnet/sdlc-kit/units/doc-style/.claude/skills/md-writing .claude/skills/md-writing $ npx degit yktsnet/sdlc-kit/units/doc-style/.claude/skills/mermaid-diagram .claude/skills/mermaid-diagram $ npx degit yktsnet/sdlc-kit/units/flow-client/.claude/skills/client-next .claude/skills/client-next $ npx degit yktsnet/sdlc-kit/units/flow-solo/.claude/skills/hand-off .claude/skills/hand-off $ npx degit yktsnet/sdlc-kit/units/flow-solo/.claude/skills/work .claude/skills/work $ npx degit yktsnet/sdlc-kit/units/flow-team/.claude/skills/issue-draft .claude/skills/issue-draft $ npx degit yktsnet/sdlc-kit/units/flow-team/.claude/skills/issue-work .claude/skills/issue-work $ npx degit yktsnet/sdlc-kit/units/flow-team/.claude/skills/rework-review .claude/skills/rework-review $ npx degit yktsnet/sdlc-kit/units/grasp/.claude/skills/grasp .claude/skills/grasp $ npx degit yktsnet/sdlc-kit/units/phase-guarantee/.claude/skills/guarantee-audit .claude/skills/guarantee-audit $ npx degit yktsnet/sdlc-kit/units/phase-mvp/.claude/skills/mvp-docs .claude/skills/mvp-docs $ npx degit yktsnet/sdlc-kit/units/walkthrough/.claude/skills/walkthrough .claude/skills/walkthrough
$ curl -fsSL --create-dirs -o .claude/agents/jp-proofreader.md https://raw.githubusercontent.com/yktsnet/sdlc-kit/main/units/doc-style/.claude/agents/jp-proofreader.md $ curl -fsSL --create-dirs -o .claude/agents/md-reviewer.md https://raw.githubusercontent.com/yktsnet/sdlc-kit/main/units/doc-style/.claude/agents/md-reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/completion-verifier.md https://raw.githubusercontent.com/yktsnet/sdlc-kit/main/units/flow-solo/.claude/agents/completion-verifier.md $ curl -fsSL --create-dirs -o .claude/agents/issue-verifier.md https://raw.githubusercontent.com/yktsnet/sdlc-kit/main/units/flow-team/.claude/agents/issue-verifier.md $ curl -fsSL --create-dirs -o .claude/agents/screen-operator.md https://raw.githubusercontent.com/yktsnet/sdlc-kit/main/units/flow-team/.claude/agents/screen-operator.md
This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 16.2k rigs:
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(./.env.*)",
"Read(**/.env)",
"Read(~/.ssh/**)",
"Bash(git push --force:*)",
"Read(**/*.pem)",
"Read(.env)",
"Read(~/.aws/**)",
"Read(**/*.key)",
"Read(**/.env.*)",
"Bash(git push --force*)",
"Bash(rm -rf:*)",
"Bash(rm -rf *)",
"Bash(sudo *)",
"Bash(sudo:*)",
"Bash(rm -rf /)",
"Bash(git push -f:*)",
"Read(.env.*)",
"Bash(git reset --hard*)",
"Bash(rm -rf /*)"
],
"ask": [
"Bash(git push:*)",
"Bash(git push *)",
"Bash(gh pr merge *)",
"Bash(git commit:*)",
"Bash(rm *)",
"Bash(gh pr merge:*)",
"Bash(gh pr create *)",
"Bash(git reset --hard:*)",
"Bash(npm publish:*)",
"Bash(git rebase *)"
]
}
} Skills (16)
qa-collectqa-senddaily-reportjp-writingmd-writingmermaid-diagramclient-nexthand-offworkissue-draftissue-workrework-reviewgraspguarantee-auditmvp-docswalkthrough
Subagents (5)
| jp-proofreader | 日本語の Markdown 文書の表現だけを校正する。書き手とは別のコンテキストで規範を当てるための役。日本語の .md を書き換えたら、コミットの前に書き手のセッションが呼ぶ。Issue / PR の本文とコード中のコメントは対象外。 |
| md-reviewer | Markdown 文書の組み立て(目的・記録との区別・重複・リンク・入口)を規範に照らして点検し、指摘だけを返す。直さない。新しい .md を作ったか、見出しを足した・消した・動かしたときに、コミットの前に書き手のセッションが jp-proofreader と並べて呼ぶ。文の表現は jp-proofreader が見る。 |
| completion-verifier | 作業ブランチを、work が宣言した対象・完了条件・保証だけを基準に確かめる。完了条件ごとに自分で実行して合否と証拠を返し、証拠にしたテストの期待値を自分で出し直す。hand-off が PR を立てる前に呼ぶ。 |
| issue-verifier | 実行者のブランチを、GitHub Issue だけを基準に確かめる。確認の項目と保証節を自分で動かし、対象と範囲を差分と突き合わせて判定し、証拠にしたテストの期待値を自分で出し直して確かめ、合否と証拠を返す。issue-work のコミットのあとに実行者が呼ぶ。 |
| screen-operator | 立ち上がっている画面をヘッドレスのブラウザで操作し、渡された道順のとおりに進めて、各手順で見えたものと画面写真を返す。合否は判定しない。issue-work で、画面に出る変更を user に見せる前に実行者が呼ぶ。 |
Similar rigs
yktsnet/dotfiles-public
An AI-agent development environment as code: Nix config plus Claude Code deny rules, hooks and skills that enforce the rules.
Fort Knox 2.1k tok · · Q48
Bastioned-successor320/learn-nanobot
Learn Nanobot with a beginner-friendly guide to AI Agent basics, interview prep, and core concepts for job seekers
Minimalist 240 tok · · Q0
mb-89/quackitect
Your agent fills, you decide. A gate ledger for spec-driven engineering: every decision arrives as a one-tap brief - on screen or phone - and nothing ships without your approval. Dependency-free Go, AGENTS.md-native.
Pragmatist 296 tok · · Q15
marnickvda/know-it-all
Agent Skills to understand a code change's design, then review its PR: walkthrough explains it in the order it runs, quiz-me tests you on it (mostly "why X over Y?"), and design-review turns what stayed unclear into design-level PR comments
Minimalist 255 tok · · Q0