mattpocock-skills

skills

Agent skills for real engineering — fork of mattpocock/skills plus personal extras (Herdr orchestrator).

Repo: https://github.com/eszxcvfd/skills

Use this README to pull the set into any other project.

---

Install vào project khác

Chọn một cách. Làm trong thư mục project đích (app/repo bạn đang code).

A. skills.sh (nhanh — copy skill vào project)

cd /path/to/your-project
npx skills@latest add eszxcvfd/skills

• Chọn skill cần dùng và agent (Claude Code, Codex, OpenCode, Cursor, …).
• Nên tick setup-matt-pocock-skills.
• Trong agent, chạy /setup-matt-pocock-skills một lần/repo (issue tracker, triage labels, docs layout).

Cập nhật sau này: chạy lại cùng lệnh npx skills@latest add eszxcvfd/skills.

B. Clone + symlink global (một bản, nhiều project)

Giữ một clone; mọi agent đọc skill qua symlink home:

git clone https://github.com/eszxcvfd/skills.git ~/src/skills
cd ~/src/skills
./scripts/link-skills.sh

Script link toàn bộ skill (trừ deprecated/) vào:

• ~/.claude/skills — Claude Code
• ~/.agents/skills — Codex / Agent Skills standard

OpenCode / pi: trỏ skill path tới clone, hoặc copy/symlink thêm:

# OpenCode user config example — skills.paths
# "~/.config/opencode/opencode.json" → "skills": { "paths": ["~/src/skills/skills"] }

Cập nhật:

cd ~/src/skills && git pull && ./scripts/link-skills.sh

C. Git submodule / subtree (skill nằm trong monorepo)

cd /path/to/your-project
git submodule add https://github.com/eszxcvfd/skills.git vendor/skills
# hoặc sparse: chỉ lấy skills/engineering + skills/productivity

Rồi cấu hình agent trỏ vào vendor/skills/skills (hoặc copy promoted buckets vào .agents/skills / .claude/skills của project).

D. Claude Code plugin (upstream managed bundle)

Plugin chính thức vẫn từ upstream Matt (read-only, auto-update theo release của ông ấy):

claude plugin marketplace add mattpocock/skills
claude plugin install mattpocock-skills@mattpocock

Dùng A hoặc B nếu bạn muốn bản fork này (có orchestrator-herdr và chỉnh sửa riêng).

---

Sau khi cài — checklist 1 project

1. /setup-matt-pocock-skills trong agent (tracker + labels + docs).
2. Router: /ask-matt khi không chắc skill nào.
3. Flow thường: /grill-with-docs → /to-spec → /to-tickets → /implement.
4. (Tuỳ chọn) Herdr multi-agent: cài Herdr, integration opencode (và agent bạn đang dùng), rồi /orchestrator-herdr — coding agent điều phối, spawn OpenCode workers theo skill project.

---

Layout repo

skills/
  engineering/   # promoted — code work (install vào project)
  productivity/  # promoted — workflow
  personal/      # setup riêng — không ship plugin
  misc/          # hiếm dùng
  in-progress/   # draft
  deprecated/    # bỏ

Promoted = an toàn copy sang project khác (kể cả orchestrator-herdr).

---

Why these skills exist

Upstream framing (Matt Pocock) — common agent failure modes and the skills that fix them:

#1: The Agent Didn't Do What I Want

"No-one knows exactly what they want"

David Thomas & Andrew Hunt, The Pragmatic Programmer

The Problem. The most common failure mode in software development is misalignment. You think the dev knows what you want. Then you see what they've built - and you realize it didn't understand you at all.

This is just the same in the AI age. There is a communication gap between you and the agent. The fix for this is a grilling session - getting the agent to ask you detailed questions about what you're building.

The Fix is to use:

• /grill-me - for non-code uses
• /grill-with-docs - same as /grill-me, but adds more goodies (see below)

These are my most popular skills. They help you align with the agent before you get started, and think deeply about the change you're making. Use them every time you want to make a change.

#2: The Agent Is Way Too Verbose

With a ubiquitous language, conversations among developers and expressions of the code are all derived from the same domain model.

Eric Evans, Domain-Driven-Design

The Problem: At the start of a project, devs and the people they're building the software for (the domain experts) are usually speaking different languages.

I felt the same tension with my agents. Agents are usually dropped into a project and asked to figure out the jargon as they go. So they use 20 words where 1 will do.

The Fix for this is a shared language. It's a document that helps agents decode the jargon used in the project.