claude-workerlisted

# Claude Worker An unofficial bridge to the Claude Code CLI installed on this machine. It is not affiliated with, endorsed by, or supported by Anthropic or OpenAI. Use this skill when the engineering leader explicitly delegates a code behavior change, new feature, bug fix, or cross-file implementation to Claude Code. It is a delegation bridge, not a code reviewer: an independent reviewer remains responsible for reviewing the diff. ## Running a task Before starting, send Claude a compact task envelope containing the goal, scope, constraints, acceptance criteria, known evidence, and required verification. Require an absolute workspace `cwd` and exactly one of task text or an absolute task file. ```text node <plugin>/scripts/claude-worker.js check node <plugin>/scripts/claude-worker.js run --cwd /absolute/workspace --task "..." node <plugin>/scripts/claude-worker.js status [<job-id>] node <plugin>/scripts/claude-worker.js result <job-id> node <plugin>/scripts/claude-worker.js resume <job-id> --task "..." node <plugin>/scripts/claude-worker.js cancel <job-id> ``` `check` reports whether the Claude Code CLI is reachable, the job store path, and the resolved model and permission mode. ## Defaults and settings Defaults are Claude Code `opus` with permission mode `auto`, stream JSON input/output, verbose output, `--safe-mode`, and explicit deny rules for git push, `gh pr merge`, and common deployment or publishing commands. Model and permission mode resolve in this order: th