菜单
Use when switching the local Docker Model Runner model on <MAC_HOST> (:12434). Handles loading/unloading models, checking what's running, and configuring Cline/opencode to use a specific local model. Only one model can run at a time on Metal.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | docker-switch |
| description | Use when switching the local Docker Model Runner model on <MAC_HOST> (:12434). Handles loading/unloading models, checking what's running, and configuring Cline/opencode to use a specific local model. Only one model can run at a time on Metal. |
| category | execution |
| runtimes | ["claude"] |
| pii_safe | true |
Docker Model Runner on <MAC_HOST> (:12434). OpenAI-compatible. One model at a time — loading a second crashes Metal compute.
From any machine (<WORKSTATION_HOST>, <INFERENCE_HOST>):
ssh <MAC_HOST> "docker-switch <alias>"
From Mac directly:
docker-switch <alias>
| Alias | Model ID | Size | Best For |
|---|---|---|---|
coder-next | docker.io/ai/qwen3-coder-next:latest | 45GB | Planning, architecture (71.3% SWE-bench) |
qwen36 | docker.io/ai/qwen3.6:latest | 21GB | Code generation (68.2% SWE-bench) |
devstral | docker.io/ai/devstral-small-2:latest | 14GB | Agentic coding (68% SWE-bench, fast) |
gemma4 | docker.io/ai/gemma4:26B | 16GB | Vision/multimodal |
nemotron | docker.io/ai/nemotron3:latest | 22GB | Long context (1M tokens) |
qwen3coder | docker.io/ai/qwen3-coder:30B-A3B-UD-Q4_K_XL | 16GB | Coding MoE |
ssh <MAC_HOST> "docker-switch list"
ssh <MAC_HOST> "docker-switch coder-next"ssh <MAC_HOST> "docker-switch qwen36"http://<PRIVATE_IP>:12434/v1sk-localdocker.io/ai/qwen3.6:latest)| Problem | Fix |
|---|---|
| Compute error 500 | Two models loaded — run ssh <MAC_HOST> "docker-switch <alias>" to flush and load one |
| Model not found | Use full ID or supported alias |
| Slow first response | Model loading into Metal RAM — normal for first request |
| TTL expiry | Models auto-unload after ~4min idle — just make a request to reload |
Models over ~25GB risk Metal compute errors when another model is also loaded.
Always use docker-switch (which unloads all first) rather than loading directly.
Mark a logical phase completion in haivemind. More formal than /snapshot — used at step/phase boundaries in multi-phase work. Creates a named, resumable point that /continue can restore from. Triggered by /checkpoint, "checkpoint this", "mark this as a checkpoint", or at phase boundaries in long-running tasks.
Run the Qwen Code CLI as a one-shot bounded executor in a plan/do/verify loop. A frontier model (or you) plans the next bounded step and verifies the result; qwen-do makes a local model do exactly that one step, smoke-test first, then stop. Use when you want a cheap local model to do grunt coding while a smarter orchestrator keeps it on rails.
Scan git commits for PII, credentials, API keys, and passwords before they are pushed. Automatically gates commits via a pre-commit hook. Use when installing PII scanning on a repo, manually scanning a file or diff for credentials, or when a commit is blocked by the hook and remediation is needed. Triggers on "scan for PII", "check for credentials", "install PII guard", "pii hook", or when a commit is blocked by a PII check.
Runs a full CORE-EEAT 80-item content quality audit, scoring content across 8 dimensions with weighted scoring by content type. Produces a detailed report with per-item scores, dimension analysis, and a prioritized action plan.
Deep analysis and debugging with Opus. Systematically investigate problems, identify root causes, and recommend fixes with evidence.
PR review skill. Two modes: /pr [link] = internal thoughts only, what to look for, what to test — NO GitHub action ever. /pr-comment [link] "make it so #1" = draft a GitHub PR comment in <USER_NAME>s voice, present for approval, post only on explicit "go". Passphrase + "go" are both required before any GitHub write. Without both gates, nothing leaves the building.