By angioz
Delegate tasks to OpenCode for token-efficient codebase exploration, search, and docs updates
Cancel a running OpenCode background job. Optionally pass a job ID; defaults to the most recent running job.
Get the output from a completed OpenCode background job. Optionally pass a job ID; defaults to the most recent job.
Delegate a task to OpenCode. Add --background for async execution. Use for codebase search, file analysis, and docs updates.
Interactive setup wizard — authenticate a provider and choose a model for OpenCode. Run this before your first /opencode:run.
Show status of OpenCode background jobs. Optionally pass a job ID to check a specific job.
Delegate tasks to OpenCode from inside Claude Code — for token-efficient codebase exploration, file analysis, and docs updates.
When Claude Code's context window fills up on large codebases, offload heavy tasks to a separate OpenCode session. Results come back as a summary without consuming your main context.
| Command | Description |
|---|---|
/opencode:setup | Interactive wizard — choose provider, authenticate in terminal, select model from live CLI output |
/opencode:run [task] | Delegate a task to OpenCode (sync by default, or --background) |
/opencode:status [job-id] | Check status of background jobs |
/opencode:result [job-id] | Get output from a completed background job |
/opencode:cancel [job-id] | Cancel a running background job |
Step 1 — Install OpenCode:
npm install -g opencode-ai
Step 2 — Add the marketplace and install the plugin in Claude Code:
/plugin marketplace add Angioz/opencode-plugin-cc
/plugin install opencode@angioz-opencode
/reload-plugins
Step 3 — Run the setup wizard:
/opencode:setup
The wizard guides you through provider selection, authenticates via opencode providers login in your terminal (API key never in chat), then picks a model from real CLI output.
claude --plugin-dir ./plugins/opencode
/opencode:setupInteractive wizard that walks through:
! opencode providers login -p [provider] in your terminal; API key entered interactively, never in chatopencode models [provider], parses the live output, and presents real model IDs via a widgetOPENCODE_MODEL to ~/.opencode-plugin/config.sh (model string only, no API key)Re-run any time to reconfigure.
/opencode:runDelegates a task to OpenCode synchronously — Claude Code waits for the result and summarizes it.
/opencode:run summarize the src/components/ directory structure
/opencode:run find all TODO and FIXME comments in the codebase
/opencode:run update the API docs in docs/api.md to match the current routes
Use --background for long-running tasks:
/opencode:run --background audit all API routes for missing input validation
Returns a job ID immediately. Check progress with /opencode:status.
/opencode:status/opencode:status # list all jobs
/opencode:status 1712345678-1234 # check a specific job
/opencode:result/opencode:result # result of the most recent job
/opencode:result 1712345678-1234 # result of a specific job
/opencode:cancel/opencode:cancel # cancel the most recent running job
/opencode:cancel 1712345678-1234 # cancel a specific job
/opencode:run list all exported functions in src/lib/ with their signatures
/opencode:run --background audit all API routes for missing input validation
/opencode:status
# ... continue working in Claude Code ...
/opencode:result
OPENCODE_MODEL in ~/.opencode-plugin/config.sh (chmod 600) — model string only, no credentialsopencode providers login — stored in ~/.local/share/opencode/auth.jsonopencode run -m [model] [message] for one-shot headless execution~/.opencode-jobs/API keys never pass through the Claude Code chat conversation — not as user input, not as Claude output, not as command arguments.
During /opencode:setup, you run ! opencode providers login -p [provider] in your terminal. OpenCode handles the interactive key prompt itself and stores credentials in its own secure store (~/.local/share/opencode/auth.json). The plugin only reads the model choice it saved during setup.
| Provider | Free tier | Key URL |
|---|---|---|
| OpenRouter | Yes — 26 free models | openrouter.ai/keys |
| Yes — Gemini free tier | aistudio.google.com/apikey | |
| Groq | Yes — rate-limited | console.groq.com/keys |
| Anthropic | No | console.anthropic.com/keys |
| OpenAI | No | platform.openai.com/api-keys |
| Ollama | N/A (local) | — |
MIT
Runs pre-commands
Contains inline bash commands via ! syntax
Bash prerequisite issue
Uses bash pre-commands but Bash not in allowed tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub angioz/opencode-plugin-cc --plugin opencodePersonal Claude Code + Codex dev stack: security hooks, AI-first code conventions, /security-review, /repo-map, /stack-check, portable statusline. Designed to complement other skills-based plugins, not replace them.
Claude Code workflow patterns: prompting, CLAUDE.md maintenance, multi-agent orchestration
Commands for loading context and priming Claude for specific tasks
(Alpha) Persistent memory, architectural decisions, and safety guardrails for Claude Code. Your agent starts every session with full project context — stack, decisions, patterns, safety rules, and a handoff from the previous session.
Auto-document entire codebase with inline comments and API docs
Expert guidance for OpenAI Codex CLI non-interactive mode operations - running prompts programmatically, parsing output formats, building automation scripts, and CI/CD integration