OpenCode ACP Channel Plugin
Paste the following prompt into your AI chat to install this skill:
Please follow https://skillhub.cn/install/skillhub.md to install @user_2e438cd6/opencode-acp-v5
About this skill
Problem
OpenClaw already provides a chat entry point, but delegating complex coding work to OpenCode can still feel fragmented by external scripts, throwaway sessions, and manual context handoffs. The OpenCode ACP channel plugin attaches OpenCode to an ACP channel so a current chat thread can start, retain, and steer a code-task session with less manual tooling glue.
How It Works
- Session start: use
/acp spawn opencodeto create an ACP session and bind the current chat thread to the targetOpenCodeinstance. - Continued interaction: persistent sessions support
/acp steerfor follow-up commands, avoiding repeated context setup. - Lifecycle control: the plugin covers session lifecycle management and permission configuration, which fits controlled code-agent invocation and debugging boundaries.
Notes
It acts mainly as a channel and orchestration layer, not a source of code-style, testing, or project rules. Outcomes depend on the OpenCode backend, model permissions, and repository context. If the environment lacks a usable backend, auth configuration, or permission policy, those prerequisites must be prepared before using it from a chat thread.
Use Cases
- Start a code-fix session in an OpenClaw chat thread and send follow-up instructions.
- Keep a session alive for cross-file debugging and steer it with next-change requests.
- Apply permission and lifecycle controls before running a code-agent task.
- Attach a chat thread to OpenCode and continue asking for revision results.
Best For
- OpenClaw workflow maintainers who need chat commands routed to OpenCode
- Engineering leads who need code-agent permission and session lifecycle controls
- Engineers debugging script or code tasks in OpenClaw who want multi-turn steering
- Integration engineers connecting external code tools to existing chat channels
Related Skills
A one-shot coding agent built on Claude Code CLI that runs non-interactively, supports a specified workdir, and can be monitored in the foreground or background.
Preview and confirm file sorting by extension, with recursive cleanup, ignore rules, and transactional rollback.
An engineering assistant for static HTML/CSS/JS pages, design-token extraction, IE8-compatible review, and structured delivery.
An engineering workflow for requirement analysis, scenario modeling, risk planning, quality gates, testing, and knowledge capture, with lightweight, standard, and full modes.