dyz-gongwen-skills
公文写作 Skills|AI 公文写作助手:防编造、守保密、GB/T 9704 一键排版,覆盖 22 种文书,支持 Claude Code / Codex / Cursor
翻訳待ち
Production-ready Claude Code configuration: 27 skills, 37 hooks, 5 rules. Includes Anthropic best practices, gstack integration, and conditional attention tags.
翻訳待ち
スキルShell · MarkdownGitHub @Cor-Incorporated自分でインストール所有者未確認
Paste into Claude or ChatGPT for installation instructions suited to your environment.
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
The maker's introduction consists of the line above. Read the repository README for more details. Repository README ↗
Loading the full README.
Read the full README ↗English | 日本語
A curated collection of skills, rules, and hooks for Claude Code — Anthropic's official CLI for Claude.
This repository provides a production-ready Claude Code configuration with 26 custom skills, 17 hook scripts (4 blocking + 13 advisory/infra), 6 rule sets, 6 utility scripts, and integration with third-party skill frameworks.
Design Philosophy: This project implements the principles from Harness Engineering Best Practices 2026 — deterministic quality gates via hooks, pointer-based documentation (ADR-002), and the feedback speed hierarchy (PostToolUse > pre-commit > CI > human review). Since ADR-006, the hook set is deliberately minimal: hard blocks are reserved for destructive/irreversible operations, and merge safety is delegated to GitHub branch protection + PR review rather than local gates.
git clone https://github.com/terisuke/claude-code-skills.git
cd claude-code-skills
chmod +x setup.sh
./setup.sh
Restart Claude Code after installation.
Claude Code can only use one API gateway per process. This repo ships scripts/claude-provider.sh so you can switch between the official Claude subscription (default, reliable for subagents/WebSearch) and z.ai GLM.
bash ~/.claude/scripts/claude-provider.sh status
bash ~/.claude/scripts/claude-provider.sh anthropic # default
bash ~/.claude/scripts/claude-provider.sh zai
Details: docs/runbooks/provider-switching.md.
setup.sh installs three launchers into ~/.local/bin so that Pi, dsh and the upstream OpenCode 2.x start with the
H17 common control the way the 1.x opencode and cursor-agent already do:
| command | what it does |
|---|---|
opencode2 / opencode2 run "…" | upstream @opencode/cli 2.x (~/.opencode-2/pkg, XDG dirs under ~/.opencode-2) with the 2.x directory plugin ~/.config/opencode/plugins/aidd-cc-v2 (opencode/aidd-cc-v2/index.ts). It runs the Cor fork's guardrails launcher from the main checkout (~/Developer/opencode/packages/guardrails/bin/opencode-guardrails, dev since Cor-Incorporated/opencode#343; override with OPENCODE_GUARDRAILS_BIN), which applies the guardrails profile and receives the H17 plugin through OPENCODE_GUARDRAILS_EXTRA_PLUGINS. Every call runs on a private server (--standalone), because the 2.x background service never receives the keys or the profile of the current call. After opencode2 run … returns, the launcher waits for the stop oracle and prints the run's H17 block / would-block / unavailable rows to stderr: in 2.x the plugin runs inside the server process, so the launcher is the headless visibility path (aidd-governance ADR-010). If the plugin cannot load, the launcher refuses to start (AIDD_CC_ALLOW_UNCONTROLLED=1 overrides). Keys: opencode2 auth login or the provider env var. |
pi / pi -p "…" | Pi 1.2.0 with ~/.pi/agent/extensions/aidd-cc.ts; default --provider openrouter --model google/gemini-2.5-flash. Uses pi's own credential store or OPENROUTER_API_KEY; with a Bitwarden session in the shell it injects the key through bw-secrets run (the TUI then runs under script -q /dev/null because bw-secrets pipes stdout). |
dsh "task" / dsh web / dsh --session-id <id> "…" | DeepSeek Harness 0.2.0-rc.1 (~/.dsh-aidd/pkg, dsh only + npm overrides dsh-tools / dsh-agent = 0.2.0-rc.1) with dsh/aidd-cc-native.mts. dsh 0.2.0-rc.1 has no terminal TUI (profiles acp / headless / sdk / sdk-minimal / web), so dsh offers one-shot, session continuation and the browser UI. Never add the rc.5 Claude-Code bridge: a second dsh-tools copy breaks the bash tool. |
Sources: pi/launcher/pi, dsh/launcher/dsh, opencode/launcher/opencode2. Runbook with the live procedure: aidd-governance design/ops/runbooks/h17-human-gate-20261017.md.
claude-code-skills/
├── skills/ # 25 custom skill definitions (SKILL.md + scripts + references)
├── rules/ # 7 global rule files (coding-style, git-workflow, quality, testing, delegation, hook-deployment, runbook)
├── hooks/ # 17 hook scripts (4 blocking + 13 advisory/infra); hooks/_unused/ holds 62 retired scripts, 56 via ADR-006
├── scripts/ # 6 utility scripts (Codex orchestration, provider switching, context monitoring); scripts/_unused/ holds retired review-pipeline helpers
├── setup.sh # One-command installation
├── settings.json # Template settings (sanitized, no personal paths)
└── README.md
Design decisions are recorded as ADRs in docs/adr/.
| ADR | Title | Status |
|---|---|---|
| 001 | PostToolUse Quality Loop | Accepted |
| 002 | CLAUDE.md Pointer Design Principle | Accepted |
| 003 | Feedback Speed Hierarchy | Accepted |
| 004 | Codex Large-Scale Delegation Model | Accepted |
| 005 | Plans.md → JSON Migration | Rejected |
| 006 | Minimal Safety Net — Hook Reduction | Accepted |
To add a new ADR, use the template.
This is the beginning of the README. Read the rest at the original. Read the full README ↗