🤖 Supported AI Agents, Skills Manifest & CLI Reference

L1DDER works across all major AI coding agents from a single canonical l1dder/ directory in your repository.

---

1. Terminal CLI Commands

npx l1dder init

Installs the l1dder/ directory (methodology, templates, rules, l1dder/skills.json manifest, and bundled skills) and configures native rules and slash commands for your selected AI agents.

# Interactive setup (asks for language and agents):
npx l1dder init

# Non-interactive setup for Cursor and Claude Code in English:
npx l1dder init --lang en --agents cursor,claude --yes

# Configure all supported agents at once:
npx l1dder init --lang en --agents all --yes

npx l1dder@latest update (or npx l1dder sync) — Update Entire Framework with BREAKING CHANGES Guard

When a new version of l1dder is released, run a single command:

npx l1dder@latest update

1. Changelog & BREAKING CHANGES Check: The CLI compares your installed version in l1dder/config.json against the new release, prints the changelog, and if any intermediate release contains BREAKING CHANGES, displays them prominently and asks: ❓ This update includes BREAKING CHANGES. Proceed with update? (yes/no) [y/N]: Only if you reply yes (or pass --yes) does the CLI proceed. If you reply no, zero files are modified.
2. Framework Update: updates 00-methodology.mdc, 10-conventions.mdc, 20-workflow.mdc, all bundled /l1-*, ponytail*, mattpocock/skills, templates/, user-guide/, METHODOLOGY.md, and refreshes all configured AI agent folders.
3. Post-Update AI Agent Steps (🤖): when your agent runs the update for you ("update l1dder"), the CLI prints a 🤖 AI Agent Instruction after L1DDER update block so the agent can add any new entries to protected project files (such as a new command in l1dder/rules/40-skills-router.mdc) without overwriting your custom settings.

🔒 Stack Protection: npx l1dder@latest update never overwrites your project-specific stack files (l1dder/rules/30-tech-stack.mdc, l1dder/rules/40-skills-router.mdc, l1dder/ci/*, and the "stack" array in l1dder/skills.json). Use --force only if you want to reset stack files to the default template.

npx l1dder changelog

Displays the release history, new features, and all BREAKING CHANGES.

npx l1dder@latest update --skills (or npx l1dder skills update) — Update ONLY Skills

If you want to update only the skills from a new l1dder release (/l1-*, ponytail*, mattpocock/skills, and your stack skills in l1dder/skills.json) while leaving l1dder/rules/, l1dder/templates/, and AGENTS.md/CLAUDE.md completely untouched:

npx l1dder@latest update --skills
# (or: npx l1dder@latest update skills / npx l1dder@latest skills update)

This updates only l1dder/skills/** and the skill/command folders in your configured AI agents (.cursor/skills/, .claude/skills/, .agents/skills/, etc.).

npx l1dder skills (list | add | install | update)

Manages project skills declaratively via l1dder/skills.json (package.json for AI Agent Skills):

# List all skills declared in l1dder/skills.json (l1dder, core, stack) and their install status:
npx l1dder skills list

# Add an external stack skill to "stack" in l1dder/skills.json, download it from GitHub into l1dder/skills/, and sync to all agents:
npx l1dder skills add vercel-labs/agent-skills --skill vercel-react-best-practices,vercel-composition-patterns
npx l1dder skills add anthropics/skills --skill frontend-design,webapp-testing

# Download all stack skills declared in l1dder/skills.json:
npx l1dder skills install

# Pull latest upstream versions from GitHub for both stack and bundled core skills (ponytail + mattpocock/skills):
npx l1dder skills update

npx l1dder doctor

Runs a fast, non-blocking check of your l1dder/ installation and the three docs/ control screens (docs/ideas.md, docs/backlog.md, docs/roadmap.md).

---

2. 📦 Skills Manifest l1dder/skills.json ("package.json for Skills")

Inside l1dder/ sits l1dder/skills.json, organizing all skills into 3 clear tiers:

{
  "l1dder": [
    "l1-onboard", "l1-idea", "l1-feature", "l1-planning", "l1-task", "l1-hotfix", "l1-chore", "l1-fast", "l1-evo"
  ],
  "core": [
    {
      "repo": "DietrichGebert/ponytail",
      "bundled": true,
      "skills": ["ponytail", "ponytail-review", "ponytail-audit", "ponytail-debt"]
    },
    {
      "repo": "mattpocock/skills",
      "bundled": true,
      "skills": [
        "grilling", "domain-modeling", "codebase-design", "tdd",
        "diagnosing-bugs", "code-review", "retro", "prototype", "writing-for-agents"
      ]
    }
  ],
  "stack": [
    {
      "repo": "anthropics/skills",
      "skills": ["frontend-design", "webapp-testing"]
    }
  ]
}

1. l1dder (workflow control skills): built-in slash commands (/l1-onboard, /l1-idea, /l1-feature, /l1-planning, /l1-task, /l1-hotfix, /l1-chore, /l1-fast, /l1-evo).
2. core (core engineering skills — bundled out of the box bundled: true):
   • DietrichGebert/ponytail: ponytail, ponytail-review, ponytail-audit, ponytail-debt.
   • mattpocock/skills (Model-invoked engineering primitives): grilling, domain-modeling, codebase-design, tdd, diagnosing-bugs, code-review, retro, prototype, writing-for-agents (along with all companion reference files DEEPENING.md, DESIGN-IT-TWICE.md, tests.md, mocking.md, LOGIC.md, UI.md, GLOSSARY-FORMAT.md, ADR-FORMAT.md).
   • Because they are bundled (bundled: true), npx l1dder init installs them offline in milliseconds, while npx l1dder skills update lets you pull fresh upstream updates from GitHub anytime.
3. stack (project stack skills): configured during /l1-onboard (or via npx l1dder skills add <repo> --skill <name>), stored in l1dder/skills.json, and preserved across npx l1dder sync.

---

3. Generated Files per AI Agent

AI Agent	--agents Flag	Files Generated in Your Repository
Cursor	cursor	• .cursor/rules/l1dder/*.mdc (5 rules with alwaysApply: true) • .cursor/skills/<skill>/* (all skills /l1-*, ponytail*, mattpocock/skills & stack) • AGENTS.md (managed <!-- l1dder:begin --> block)
Claude Code	claude	• CLAUDE.md (managed <!-- l1dder:begin --> block) • .claude/rules/l1dder/*.md (modular rules) • .claude/skills/<skill>/* (complete skill folders)
OpenAI Codex	codex	• AGENTS.md • .agents/skills/<skill>/* (complete folders for all skills)
Windsurf	windsurf	• .windsurf/rules/l1dder.md • .windsurf/workflows/<skill>.md (slash workflows) • AGENTS.md
GitHub Copilot	copilot	• .github/copilot-instructions.md (managed marker block) • .github/prompts/<skill>.prompt.md (prompt files)
OpenCode	opencode	• AGENTS.md • .agents/skills/<skill>/*
Cline / Roo Code	cline	• .clinerules/l1dder.md • .roo/rules/l1dder.md • AGENTS.md

💡 l1dder/skills/ is the single source of truth: npx l1dder sync copies it into the skills folders of the selected agents, so edit skills there rather than in agent folders. Cursor always reads .agents/skills/ and reads .claude/skills/ only with Third-Party Imports enabled, so cursor + claude installs both; .cursor/skills/ is used only when Cursor is the only agent with a skills folder.

🩹 Commands didn't appear in the / menu? See the step-by-step guide.