Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/dryvist/claude-code-plugins/terrakube-opsnpx skills add dryvist/claude-code-plugins --skill terrakube-opsgit clone --depth 1 https://github.com/dryvist/claude-code-pluginsWrote 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/dryvist/claude-code-plugins/terrakube-ops)<a href="https://agentmods.dev/skills/dryvist/claude-code-plugins/terrakube-ops"><img src="https://agentmods.dev/badge/skills/dryvist/claude-code-plugins/terrakube-ops.svg" alt="Measured on agentmods" 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 | $0.00096 | $0.01263 |
| Opus 5 | $0.00048 | $0.00632 |
| Sonnet 5 | $0.00019 | $0.00253 |
| Haiku 4.5 | $0.00010 | $0.00126 |
Grade A, and why
terrakube-ops 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 yesterday.
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 — 118 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Terrakube operations
Terrakube is a self-hosted remote-execution platform for OpenTofu/Terraform — plans and applies run on its own executor, not on the operator's machine or in CI. This skill is the generic operating model; your own workspace names, hostnames, and credential source stay in your own inventory.
Canonical workflow
tofu login <terrakube-hostname> # once per machine — opens a browser SSO flow
tofu plan # runs remotely; CLI streams the live plan log
tofu apply # same remote workspace, same run
tofu state list # inspect state / outputs afterward
tofu output
tofu plan/apply from the CLI and starting a run from the Terrakube web UI
both drive the same remote executor and the same job log — pick whichever is
convenient; they're interchangeable.
Local
tofu init -backend=false+tofu validatecatches syntax errors without contacting any provider — run that first, it's free and instant compared to a remote plan.
Never use a targeted apply
A targeted apply (-target=...) can leave the real infrastructure and any
downstream-published inventory describing two different worlds — whatever
consumes that inventory (configuration management, DNS, monitoring) now
disagrees with reality. Apply the reviewed whole-workspace plan. For
state surgery, prefer moved/removed blocks over -target or manual state
edits.
Workspace locking and recovery
Terrakube owns one run queue and one lock per workspace — a second run simply queues behind the active one; independent workspaces don't share a lock.
- Cancel a genuinely stuck run from the workspace UI, not by force-unlocking blind.
- Confirm the executor has actually stopped before force-cancelling or unlocking — cancelling a run that's still writing state is how corruption happens.
- Prefer fixing forward (revert the config change in git, plan/apply again) over restoring an older state version. When state genuinely is corrupted, preserve the current state version before restoring an older one — you may need to diff them later.
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.
- yesterday First seen · 118 lines · 96 tokens per session scan A f4040833decb
terrakube-ops is a skill published in the GitHub repository dryvist/claude-code-plugins (3 stars, last pushed today), licensed Apache-2.0. It adds 96 tokens to every session and 1,263 once invoked, about $0.0005 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 2026-09-04.
Other skills, from other repositories
vast-gpu
Rent, manage, and destroy GPU instances on vast.ai. Use when user says "rent gpu", "vast.ai", "rent a server", "cloud gpu", or needs on-demand GPU without owning hardware.
pinme-r2
Use when a PinMe Cloudflare Worker needs R2 object storage, including secure file or image upload, streaming download, metadata lookup, deletion, listing, Range requests, or R2+D1 coordination. Guides AI to use PinMe's automatically injected env.R2 binding without R2 credentials or manual Wrangler configuration.
api-workers
Cloudflare Workers deployment using createWorkerHandler from @cyanheads/mcp-ts-core/worker. Covers the full handler signature, binding types, CloudflareBindings extensibility, runtime compatibility guards, and wrangler.toml requirements.
cloudflare-workers-publish
Deploy static HTML files to Cloudflare Workers with 1Password credential management.
api-gateway
API Gateway patterns (Kong, Traefik, AWS API Gateway) — rate limiting, auth, routing, versioning. Use when implementing API gateway, reverse proxy, or API management.
bailian-train-deploy
用百炼 CLI (bl) 走完"数据→微调训练→导出→部署→调用"的完整闭环，或跳过训练直接部署基座模型。支持文本模型（SFT/DPO/CPT）、音频 TTS 模型（CosyVoice）、图像生成模型（Wan2.7）和视频生成模型（Wan i2v/kf2v）微调。涵盖数据集校验/上传、创建微调任务、等待训练、导出最佳 checkpoint、创建推理部署、等待就绪、给出调用示例。当用户提到在百炼 / DashScope / 阿里云模型工作室上"训练模型""微调""fine-tune""finetune""部署模型""模型上线""把微调模型跑起来/调用""训练一个推理模型""继续预训练""LoRA/SFT/DPO…