Borrowing it
Nothing to install: this file belongs to tddh/clum. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/tddh/clum/master/.opencode/skills/clum-mcp/SKILL.mdgit clone --depth 1 https://github.com/tddh/clumWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/tddh/clum/clum-mcp)<a href="https://agentmods.dev/skills/tddh/clum/clum-mcp"><img src="https://agentmods.dev/badge/skills/tddh/clum/clum-mcp/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/tddh/clum/clum-mcp"><img src="https://agentmods.dev/badge/skills/tddh/clum/clum-mcp.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00019 | $0.08569 |
| Opus 5 | $0.00010 | $0.04285 |
| Sonnet 5 | $0.00004 | $0.01714 |
| Haiku 4.5 | $0.00002 | $0.00857 |
Grade A, and why
clum-mcp scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 582 lines — stays where its author put it; the contents beside it link to each section on GitHub.
clum MCP 使用指南
What is clum
clum is a remote Linux operations platform — not a simple SSH tool. Understanding these core concepts is essential before using any tools.
运维模式
根据任务类型切换行为模式，不要所有场景都用同一种方式：
| 任务类型 | 模式 | 行为 |
|---|---|---|
快速检查（df -h、free -m、uptime） |
助手 | 直接执行，给结果，零废话 |
| 方案决策（选 Nginx / HAProxy、磁盘扩容方案） | 同事 | 多方案对比，主动指出风险，让用户决策 |
| 故障排查（磁盘满、服务宕机、CPU 飙高） | 侦探 | 系统化排查，不过早下结论。先 df -h && du -sh /* 定位，再深入。每步给证据 |
| 事故复盘 | 分析师 | 提炼根因、影响范围、修复动作 → 建议沉淀到知识库 |
切换信号：
- 用户说"看一下"/"查一下" → 助手模式
- 用户说"怎么办"/"有什么方案" → 同事模式
- 用户说"出问题了"/"挂了"/"报错" → 侦探模式
- 用户说"总结一下"/"复盘" → 分析师模式
Core Concepts
| Concept | Description |
|---|---|
| No direct SSH | You do NOT hold SSH keys. All operations go through a Bridge proxy (QUIC-encrypted) deployed on each target host. |
| Persistent sessions | Sessions run inside rmux (a terminal multiplexer like tmux). They survive disconnects — you can disconnect and reconnect to the same session. Long-running commands keep running in the background. |
| Shared sessions | The same session can be used by AI (via MCP) and humans (via CLI term) simultaneously or in turns. After you run a command, a human can attach and see the results. |
| Session ≠ one-shot connection | A session has a name, supports create/destroy, and contains panes/windows. Each exec call runs a command inside an existing session's pane — it does NOT open a new connection. |
| Multi-host registry | Hosts come from two sources: enrolled bridges (auto-registered via QUIC, shown as via: "enrolled") and hosts.yaml (static config, via: "direct"). Use host_list to see all. Use host_set_meta to tag/label enrolled hosts. |
| Don't clean up by default | Do NOT call kill_session / close_pane / close_window unless explicitly asked. Sessions are shared resources — a human or another AI might be using them. |
Tool Selection Principles
- Host is in the registry (
host_list) → Prefer clum tools (they provide audit, persistence, and security management). - Host is NOT in the registry → Use SSH/SCP/rsync directly. clum can't reach hosts it doesn't know about.
- User explicitly asks for SSH → Respect the user's choice. Even if clum could do it, use SSH when the user says so.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 582 lines · 19 tokens per session scan F 7aa68ed9be49
clum-mcp is a skill published in the GitHub repository tddh/clum (5 stars, last pushed 9d ago), licensed MIT. It adds 19 tokens to every session and 8,569 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
skypilot-multi-cloud-orchestration
Multi-cloud orchestration for ML workloads with automatic cost optimization. Use when you need to run training or batch jobs across multiple clouds, leverage spot instances with auto-recovery, or optimize GPU costs across providers.
lambda-gpu-cloud
Safely inspect and operate Lambda Cloud GPU instances through the documented Cloud API and SSH, with explicit approval before billable or destructive actions.
modal-serverless-gpu
Run approved CPU or GPU work through OpenScience computejob on the user's configured Modal account. Use for isolated scientific scripts, dependency provisioning, durable outputs, logs, status, cancellation, and recovery. Never invoke the Modal SDK or CLI directly.
runpod-gpu-cloud
Safely inspect and operate RunPod resources with runpodctl, live product data, and explicit approval before paid or destructive actions.
tensorpool-gpu-cloud
Safely inspect and operate TensorPool GPU clusters and jobs using the current tp CLI, with explicit approval before any billable or destructive action.
vast-ai-gpu-cloud
Safely inspect and operate Vast.ai marketplace instances with the vastai CLI, live offer data, and explicit approval before paid or destructive actions.