skills
來自 scenario-labs/skills 的 67 項代理技能 — 包含 scenario、scenario-game-assets、scenario-video。使用 npx skills add scenario-labs/skills 安裝。
AI 翻譯
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證Claude Code / Codex / Cursor Agent Skill: defensive Dependabot triage with lockfile parity & exposure mapping. Plus team git workflow scripts (releases, PR helpers, branch resets).
翻譯準備中
取自儲存庫 README
正在載入完整 README。 以下為已儲存的摘錄。
A personal grab-bag of git workflow tooling, shell config, Claude Code skills, SQL helpers, and an AI-pairing methodology for junior engineers. Each top-level directory is independent — clone the repo and pull in only what you want.
Pick the parts you want — none of them depend on each other.檢視完整 README ↗A personal grab-bag of git workflow tooling, shell config, Claude Code skills, SQL helpers, and an AI-pairing methodology for junior engineers. Each top-level directory is independent — clone the repo and pull in only what you want.
| Path | What it is |
|---|---|
scripts/ | Git workflow scripts (merge into ephemeral branches, reset branches, send notifications, AWS SSM helpers). |
dotfiles/ | Shell config: bashrc.overrides (aliases, git/AWS shortcuts) and repohue (per-repo terminal colours). See dotfiles/README.md. |
skills/ | Claude Code skills: dependabot-triage + dependabot-triage-py (automated CVE triage), session-loop (multi-day project pause/resume toolkit). Built from SKILL.md.tmpl templates via a {{include}} resolver, with reusable fragments under skills/_shared/. |
socratic-method/ | Makes an AI coding agent (Claude Code, Codex, etc.) act as a Socratic tutor for a junior engineer — question-first across 7 phases (pre-coding, design review, PR review, postmortems, estimation) instead of just generating code. See socratic-method/README.md. |
.claude/ and .agents/ | Claude Code project configuration and agent definitions. |
LLM Custom Instructions.md | Custom instructions / coding protocol used with LLMs (skeptical-sparring persona, phased workflow). |
idempotent_constraints.sql | Example pattern for idempotent Postgres ADD CONSTRAINT migrations. |
Pick the parts you want — none of them depend on each other.
git clone https://github.com/<your-fork>/git-practices ~/tern-work/git-practices
# Theme engine
ln -s ~/tern-work/git-practices/dotfiles/repohue.sh ~/.config/repohue.sh
# Aliases + helpers (bash)
ln -s ~/tern-work/git-practices/dotfiles/bashrc.overrides ~/rao_bashrc_overrides
echo 'if [ -f ~/rao_bashrc_overrides ]; then . ~/rao_bashrc_overrides; fi' >> ~/.bashrc
# Open a new terminal. cd into any git repo to see colours kick in.
Full details and customisation knobs: dotfiles/README.md.
Put scripts/ on your PATH:
echo 'export PATH=$PATH:$HOME/tern-work/git-practices/scripts' >> ~/.bashrc
Then use them directly: merge_into.sh <target-branch>, new_pr.sh, ssm_login.py, etc. Most scripts have inline --help or are short enough to read.
Each skill lives under skills/<name>/. Skills with a SKILL.md.tmpl are compiled to SKILL.md by resolving {{include ...}} directives against shared fragments:
scripts/build-skill.sh skills/dependabot-triage # one skill
scripts/build-all-skills.sh # every templated skill
Publish a skill to its dedicated repo (and auto-build first):
scripts/publish-skill.sh skills/dependabot-triage
scripts/publish-skill.sh skills/dependabot-triage-py
scripts/publish-skill.sh skills/session-loop
Read socratic-method/README.md first — status
and which phases have seen less real-world use. Setup (Claude Code, Codex, or
any AGENTS.md-reading tool) is covered there in full:
socratic-method/README.md#setup.
MIT — see LICENSE.