Make CLI Agents Follow the Same Method
How do you leave rules and steps for the next run?
Write project instructions, skills, and reusable workflows.
📌 Learning goals
- Explain project instructions, skills, and one-off prompts in your own words.
- Write the rule card with four sections — purpose, prohibitions, verify commands, delivery format — that the agent actually reads.
- Build a read-only review skill; after testing, git status shows no unexpected changes.
- Know that a "shared core" does not mean identical filenames and permissions across CLIs.
Entry conditions
Complete A1's read-only and small-change loop; the same demo repo keeps working.
🧭 Lessons on this site
Read in the suggested order; checkboxes share the same browser progress as the /learn track pages.- 01Your First AI Task: How to Give Instructions It Gets Right
Why first tasks usually fail — not because the tool is weak, but because the instruction has no output, no input, and no destination. Three elements, three real examples (each with its vague version for contrast), and how to judge how much freedom to give.
7 min - 02Config Files, Dotfiles, and Package Managers: What Are Those Files Starting with a Dot?
Why every tool asks you to edit a file starting with a dot, the difference between YAML and JSON, why wrong indentation breaks everything, what problems package managers like brew and apt solve, and how much regex you actually need to know.
11 min - 03Complete Guide to the Slash Command System
From trigger points to real-world use: an in-depth look at the design philosophy behind 30+ slash commands, the reasons they were created, practical scenarios, and how to use this system to greatly boost work efficiency.
14 min - 04A Complete Tutorial for Matt Pocock Skills: The 115K Star AI Coding Engineering Discipline System, a Full Walkthrough of 15 Skills from Installation to Practice
mattpocock/skills: 35 Agent Skills released by TypeScript legend Matt Pocock from his own .claude directory, with 115K+ Stars, 2.7M+ installs, and 60K newsletter subscribers. This is not vibe coding; it is an AI coding discipline system condensing decades of software engineering experience.
24 min - 05In-Depth Analysis of Addy Osmani's agent-skills Open Source Project
A complete introduction to Google Chrome Engineering Director Addy Osmani's agent-skills project: the design philosophy behind 21 AI-assisted development skills, real-world applications, and how it inspired the upgrade of our slash command system.
10 min - 06A Comprehensive Classification Matrix for the OpenClaw Skill Library: 227 Skills × Four Categories × a Full Business-Process Map
A complete classification matrix of the 227 installed skills in our OpenClaw system. Reorganized into four categories: everyday core tools, finance-related, coding-related, and other, with each category further arranged into 10 business stages. Includes skill usage analysis, overlap detection, and trimming recommendations.
13 min
🎯 Curated resources
| Resource | Who it's for | Priority | Why |
|---|---|---|---|
Official docs Claude Code — Memory（CLAUDE.md） | Writing persistent rules for Claude Code | ⭐⭐⭐⭐⭐ | The official word on what belongs in rules and what moves to references. |
Official docs Claude Code — Skills | Building your first skill | ⭐⭐⭐⭐⭐ | SKILL.md structure, loading locations, and permissions. |
Official docs Gemini CLI — GEMINI.md | Project context for Gemini CLI | ⭐⭐⭐⭐ | Directory scope and loading behavior. |
Official docs Codex — AGENTS.md | Using Codex / ChatGPT Build | ⭐⭐⭐⭐ | How rules load, merge, and fall back. |
Official docs OpenCode — Rules & Skills | Using OpenCode | ⭐⭐⭐⭐ | AGENTS.md-first rules and skill locations. |
Shared standard Agent Skills Specification | Making core content portable | ⭐⭐⭐⭐ | The minimal shared format with readable examples. |
Examples mattpocock/skills | Comparing writing styles | ⭐⭐⭐⭐ | Short skills used in real engineering work. |
Examples obra/superpowers | After your first skill | ⭐⭐⭐⭐ | See how real workflows decompose into skills. |
Context tools yamadashy/repomix | Feeding repo content to an agent | ⭐⭐⭐⭐⭐ | Generates a one-shot codebase snapshot. |
Context tools langchain-ai/openwiki | Large repos needing on-demand docs | ⭐⭐⭐⭐ | Builds a continuously updatable repo wiki. |
Practice anthropics/prompt-eng-interactive-tutorial | When CLI-8 stalls on the shared core | ⭐⭐⭐⭐ | Compare prompt styles step by step. |
🛠 Hands-on practice (upstream)
Full exercises & starter codeSummaries from the upstream curriculum; full code, cost, and latency estimates live upstream.
- CLI-5: write the project rule card — only purpose, prohibitions, verify commands, and delivery format; move long references to docs/.
- CLI-6: build a review skill — read-only diff review returning PASS or a list of gaps; it never commits or pushes by itself.
- CLI-7: cross-tool mapping — take the same rules and skill to another CLI and re-map its accepted frontmatter, permissions, and tool names.
- CLI-8: compare prompt styles — when stuck on the shared core, work through the official prompt tutorial step by step.
- Common fix: if the agent skips your many rules, cut the backstory and keep observable behavior; mandatory safety checks belong in hooks/policies, not prose reminders.
✅ Self-check
- I can explain project rules, skills, and one-off prompts in my own words.
- My rule card has purpose, prohibitions, verify commands, and delivery format — and the agent reads it.
- My review skill only reads changes; after testing, git status --short shows nothing unexpected.
- I know a "shared core" does not mean identical filenames and permissions across CLIs.
Adapted from awesome-agentic-ai-zh (MIT, by Wenyu Chiou) v2026.09.23; links checked 2026-08-27. Stars mark learning priority (⭐⭐⭐⭐⭐ = you will get stuck without it), not popularity. MIT License · Curriculum structure last updated 2026-10-03. Content is still being filled in; lessons marked “in progress” are not live yet.