№ 0891Skill
Muse plugin for Claude Code
Port of openai/codex-plugin-cc that lets Claude Code call Muse for reviews or delegated tasks via muse exec, with /muse:review, /muse:adversarial-review and background job commands.
# Muse plugin for Claude Code Use Muse from inside Claude Code for code reviews or to delegate tasks to Muse. Port of [openai/codex-plugin-cc](https://github.com/openai/codex-plugin-cc): same command surface and architecture, but the worker is the Muse CLI (`muse exec`) instead of the Codex CLI. No app-server, no extra services — the companion script shells out to `muse exec` and tracks background jobs in local state files. ## What You Get - `/muse:review` for a normal read-only Muse review - `/muse:adversarial-review` for a steerable challenge review - `/muse:rescue`, `/muse:transfer`, `/muse:status`, `/muse:result`, and `/muse:cancel` to delegate work, hand off sessions, and manage background jobs ## Requirements - **Muse CLI installed and on PATH** (this repo's host already ships it as `muse`). Run `/muse:setup` to verify. If the first delegated task fails with an auth error, run `muse login` or `muse auth set`. - **Node.js 18.18 or later** (companion runtime only, no dependencies). ## Install Add the marketplace in Claude Code: ```bash /plugin marketplace add <your-fork-or-path>/muse-plugin-cc ``` Install the plugin: ```bash /plugin install muse@muse-plugin-cc ``` Then run: ```bash /muse:setup ``` One simple first run is: ```bash /muse:review --background /muse:status /muse:result ``` ## Usage ### `/muse:review` Runs a normal Muse review on your current work (read-only, never edits). ```bash /muse:review /muse:review --base main /muse:review --background ``` Supports `--base <ref>`, `--scope auto|working-tree|branch`, `--model <id>`, `--effort <level>`, `--wait`, `--background`. Takes no focus text — use `/muse:adversarial-review` to steer. ### `/muse:adversarial-review` Steerable review that questions the implementation and design: ```bash /muse:adversarial-review /muse:adversarial-review --base main challenge whether this was the right caching and retry design /muse:adversarial-review --background look for race conditions and question the chosen approach ``` Same target selection and flags as `/muse:review`, plus free focus text. ### `/muse:rescue` Hands a task to Muse through the `muse:muse-rescue` subagent (write-capable by default): ```bash /muse:rescue investigate why login fails on staging /muse:rescue --background migrate the billing module to the new API /muse:rescue --resume apply the top fix /muse:rescue --read-o



ChatForm
Tgmlabs