Code Agent Orchestration (OCA)

Skill for orchestrating coding agent sessions from OpenClaw, including OCA requests such as "let oca do...", "ask oca to...", or "have oca handle...". Covers launching, monitoring, plan approval, lifecycle management, and worktree decisions.

49 stars 17 forksAI, LLMs & Agents

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add goldmar/openclaw-code-agent --skill Code Agent Orchestration (OCA)
Manual install
git clone --depth 1 https://github.com/goldmar/openclaw-code-agent.git /tmp/openclaw-code-agent
mkdir -p ~/.claude/skills
cp -r /tmp/openclaw-code-agent/skills/code-agent-orchestration ~/.claude/skills/Code-Agent-Orchestration-OCA-

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Code Agent Orchestration

`openclaw-code-agent` runs Claude Code, Codex, or OpenCode sessions as background coding jobs. `oca` is the built-in short name for OpenClaw Code Agent ("let oca do …", "ask oca to …", "have oca handle …").

## Launch and continue

Use `ocaSessionId` for public tool session references, preferably the exact OCA ID from the launch receipt. `session` is a deprecated input alias; if both are supplied they must match exactly. Never repair a missing or masked reference by choosing the newest session.

- `agent_launch(prompt, name, workdir)`: give a short kebab-case `name` and the repository as `workdir`. Chat routing is automatic.
- Defaults (plugin config): `permission_mode: "plan"` (the agent plans first), `plan_approval: "delegate"` (you review the plan), `worktree_strategy: "delegate"` (the agent works on a branch; you decide what happens to it).
- Continue a session with `agent_respond(ocaSessionId, message)`; it also resumes a stopped, suspended or completed one. Launch again only for independent work, never from a wake about the same task.
- Codex mid-turn: `agent_respond` steers the running turn (`interrupt=true` restarts it). A wrong turn: `agent_launch(resume_session_id, fork_session=true, rewind_turns=1, prompt)`. Context full: `agent_session_action(action="compact")`; a review before merging: `action="review"`.
Read the full skill on GitHub

Details

Skill name
Code Agent Orchestration (OCA)
Path
skills/code-agent-orchestration/SKILL.md
License
MIT
Repo updated
17 November 2026

FAQ

What does the Code Agent Orchestration (OCA) skill do?

Skill for orchestrating coding agent sessions from OpenClaw, including OCA requests such as "let oca do...", "ask oca to...", or "have oca handle...". Covers launching, monitoring, plan approval, lifecycle management, and worktree decisions.

How do I install the Code Agent Orchestration (OCA) skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add goldmar/openclaw-code-agent --skill Code Agent Orchestration (OCA). Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Code Agent Orchestration (OCA) popular?

It lives in goldmar/openclaw-code-agent, which has 49 GitHub stars and 17 forks.

Related skills

Prompt Architect

114

DylXBT/claude-orchestrator-skills

Turn a vague request into a reliable, reusable prompt with role, context, task, constraints, and examples. Use when the user keeps re-rolling outputs to get something usable, or wants a repeatable prompt for a task they do often — in any field.

AI, LLMs & Agents

Vane Search

113

szsip239/teamclaw

Vane (Perplexica) AI 搜索引擎 API 调用。通过本地部署的 Vane 服务进行带引用的深度搜索,支持网页/学术/Reddit 来源,三档深度模式(speed/balanced/quality),返回 AI 总结 + 来源引用。适用于调研、信息收集、深度分析等场景。

AI, LLMs & Agents17

Agentation

79

bamzc/claude-skills-frontend

Add Agentation visual feedback toolbar to any React project

AI, LLMs & Agents10

Self Improving + Proactive Agent

113

szsip239/teamclaw

Self-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when (1) a command, tool, API, or operation fails; (2) the user corrects you or rejects your work; (3) you realize your knowledge is outdated or incorrect; (4) you discover a better approach; (5) the user explicitly installs or references the skill for the current task.

AI, LLMs & Agents17

Humanizer

70

alexdcd/Mafia-Claude-Skills

Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases. Credits: Original skill by @blader - https://github.com/blader/humanizer

AI, LLMs & Agents18

Multi Agent Cn

113

szsip239/teamclaw

通用多Agent调度系统(中文版):将主Agent变为纯调度员,所有任务通过 sessions_spawn 委派给5个持久化子Agent。支持轮询调度、先回复再派遣协议、 sessionKey固定复用。用户可自定义调度员角色和子Agent名称/人设。

AI, LLMs & Agents17

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.