All Agents
OpenClaw skill (gamora-approved)
Choose how to use this skill
You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.
1. Native installer
This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.
Do not guess an installer command or replace an existing version without reviewing the diff.
2. Complete package recommended
Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.
No complete ProSkills package is published for this listing yet.3. Prompt-only
Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.
Need only the instruction file?
Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.
No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.
// RATINGS
Not yet listed on ClawHub or SkillsMP
// README
all-agents
If you use Claude Code across multiple projects, you're repeating yourself — same CLAUDE.md, same workflows, same prompts copied everywhere. all-agents gives you a single source of truth: one config, every project inherits it.
It includes aaa, a CLI toolkit with autonomous planning (Ralph), parallel code review, web research, and push notifications — plus 18 skills and 9 slash commands for Claude Code.
Prerequisites
- Bun (required)
- Claude Code (required for full value — skills, agents, slash commands)
- gh CLI (optional, for GitHub search)
Quickstart
git clone https://github.com/otrebu/all-agents ~/dev/all-agents
cd ~/dev/all-agents/tools && bun install
bun run dev setup --user
# Now try it in any project:
cd ~/your-project
aaa setup --project # Link shared config
aaa review --base main --dry-run # Find issues in your changes (11 parallel reviewers)
command not found: bun — Install: curl -fsSL https://bun.sh/install | bash
command not found: aaa — Add ~/.local/bin to PATH: echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.zshrc
What You Get
| What | Command | Description |
|---|---|---|
| Plan features with AI | aaa ralph plan vision | Interactive planning — vision, roadmap, stories, subtasks |
| Build autonomously | aaa ralph build | Agents implement subtasks one by one with validation gates |
| Refresh model registry | aaa ralph refresh-models | Discover provider models and update the dynamic registry |
| Review code (11 agents) | aaa review --headless | Security, data integrity, performance, accessibility, and more |
| Research the web | aaa parallel-search | Multi-angle web research with up to 30K chars/result |
| Get notified | aaa notify "Done" | Push notifications when Claude finishes work |
| Search GitHub | aaa gh-search "react hooks" | Find real-world code examples with intent-based ranking |
| Sync context | aaa sync-context --target ~/project | Share docs across projects with watch mode |
Ralph is an autonomous dev framework: you define a vision, and agents plan, build, review, and calibrate — while you stay on the loop, not in it. Each iteration is memoryless (fresh context, no drift). Full docs: docs/ralph/README.md
aaa ralph build --provider opencode --model openai/gpt-5.3-codex
aaa ralph build --provider cursor
aaa ralph build --provider codex --model gpt-5.3-codex
Full CLI reference: tools/README.md
Setup
| Global (Recommended) | Project-Level | |
|---|---|---|
| Use when | You want all-agents everywhere | You want it in one specific project |
| What it does | All Claude Code sessions inherit config | Just that project gets shared docs |
| Command | bun run dev setup --user | aaa setup --project |
Global User Config
git clone https://github.com/otrebu/all-agents ~/dev/all-agents
cd ~/dev/all-agents/tools
bun install
bun run dev setup --user
The setup wizard builds the CLI binary, creates ~/.local/bin/aaa, checks your PATH, and prompts to set CLAUDE_CONFIG_DIR.
Result: All Claude Code sessions use this repo's commands, agents, and skills.
Note:
CLAUDE_CONFIG_DIRis not officially documented. Claude may still create.claude/settings.local.jsonin workspaces.
Project-Level Setup
cd your-project
aaa setup --project
Links context/ to all-agents shared docs, creates docs/planning/ and docs/research/ directories, and checks CLAUDE_CONFIG_DIR.
Result: Project gets shared standards via context/, keeps local planning/research in docs/.
Using with Claude Code
Claude Code extends with three mechanisms:
- Slash Commands (
/name) — Type in chat to trigger. Example:/dev:git-commit - Skills — Auto-detected from your request. Example: "what did I work on today?" triggers
dev-work-summary - Sub-agents — Background workers spawned by skills. Rarely invoked directly.
Stability: stable = battle-tested | beta = works, may evolve | experimental = may break
| Command | Description | Stability |
|---|---|---|
/dev:git-commit | Create conventional commits | stable |
/dev:git-multiple-commits | Create multiple commits | stable |
/dev:start-feature | Create/switch feature branches | stable |
/dev:complete-feature | Merge feature to main | stable |
/dev:consistency-check | Verify docs match code, find contradictions | beta |
/parallel-search | Multi-angle web research | beta |
/context:plan-multi-agent | Plan docs with Opus agents | experimental |
/meta:claude-code:create-skill | Create a new skill | beta |
/meta:claude-code:create-agent | Create a sub-agent | beta |
/meta:create-cursor-rule | Create .cursorrules file | experimental |
| Agent | Description | Stability |
|---|---|---|
atomic-doc-creator | Create missing atomic documentation | experimental |
parallel-search | Multi-angle web research | beta |
subtask-reviewer | Review subtasks using vertical slice test | experimental |
task-generator | Generate technical tasks from stories | experimental |
accessibility-reviewer | WCAG, keyboard nav, ARIA, color contrast | experimental |
data-integrity-reviewer | Null checks, race conditions, schema violations | experimental |
dependency-reviewer | Version compat, licenses, circular deps | experimental |
documentation-reviewer | Docstrings, API docs, README gaps | experimental |
error-handling-reviewer | Swallowed exceptions, missing catch, async issues | experimental |
intent-alignment-reviewer | Code matches specification | experimental |
maintainability-reviewer | Coupling, naming, DRY, SRP issues | experimental |
over-engineering-reviewer | YAGNI, premature abstraction | experimental |
performance-reviewer | O(n^2), memory leaks, N+1 queries | experimental |
security-reviewer | OWASP Top 10, injection, XSS, auth, secrets | experimental |
test-coverage-reviewer | Missing tests, untested branches, assertions | experimental |
synthesizer | Aggregate and dedupe findings from reviewers | experimental |
triage | Curate findings: select must-review, group by root cause, filter noise | experimental |
| Skill | Description | Stability |
|---|---|---|
aaa-feature-wizard | Wizard for adding new aaa CLI commands | experimental |
brainwriting | 5 parallel idea explorations, then synthesize | beta |
context-atomic-doc | Create/update atomic docs (blocks, foundations, stacks) | experimental |
dev-work-summary | Scan ~/dev for today's git work | beta |
doc-analyze | Tiered doc analysis (T1: Haiku lookup, T2: Sonnet gaps, T3: Opus deep) | experimental |
gh-search | Search GitHub for code examples and patterns | experimental |
interrogate-on-changes | Surface decisions, alternatives, confidence | experimental |
code-review | Orchestrate 11 reviewers in parallel | experimental |
ralph-build | Autonomous build loop for subtasks | experimental |
ralph-calibrate | Check intention drift, technical quality | experimental |
ralph-plan | Interactive vision planning (Socratic method) | experimental |
ralph-review | Review auto-generated planning artifacts | experimental |
ralph-status | Display build progress and stats | experimental |
ralph-prompt-audit | Audit ralph command examples in prompts vs live CLI | experimental |
run-guide-and-fix | Walk through GUIDE.md with agent-browser, fix bugs | experimental |
write-guide | Generate comprehensive GUIDE. |