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/miaodx/roboclaws/cloudml-eval-opsnpx skills add MiaoDX/roboclaws --skill cloudml-eval-opsgit clone --depth 1 https://github.com/MiaoDX/roboclawsWrote 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/miaodx/roboclaws/cloudml-eval-ops)<a href="https://agentmods.dev/skills/miaodx/roboclaws/cloudml-eval-ops"><img src="https://agentmods.dev/badge/skills/miaodx/roboclaws/cloudml-eval-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 | $0.00099 | $0.03923 |
| Opus 5 | $0.00049 | $0.01962 |
| Sonnet 5 | $0.00020 | $0.00785 |
| Haiku 4.5 | $0.00010 | $0.00392 |
Grade A, and why
cloudml-eval-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 4d 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 — 328 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CloudML Eval Ops
Run CloudML as an execution environment for an existing Roboclaws Eval Harness manifest. Do not make CloudML a product backend and do not reimplement CloudML or storage APIs in Roboclaws.
Ownership
- Let Roboclaws own row selection, commands, graders, result schemas, and final
reports. Read
../eval-harness/SKILL.mdbefore selecting rows. - Use the installed
cml-shared,cml-resource, andcml-trainskills and the officialcmlCLI for context, resources, YAML, submit, describe, logs, events, stop, and task status. - Use the installed
executorskill only for cross-platform Repo and JuiceFS/FDS operations. Read its CloudML defaults before those operations. - Keep cluster, queue, image, mount, endpoint, and credential values in existing environment-owned CML/executor configuration or run-local ignored files. Do not add them to this skill, tracked source, commands printed in reports, or normal eval artifacts.
Task Authorization
Treat an explicit human request to run or refresh a repo-scoped CloudML eval as authorization for all necessary in-scope operations: read-only preflight, source and asset packaging, scoped upload, image publication when required by the selected row, task submission, monitoring, logs/events, result download, collection, stopping a failed or stalled task, and repair followed by a new attempt. Do not pause for per-command or per-stage approval while workspace, provider routes, queue/resource class, maximum concurrency, and documented cost envelope remain unchanged.
Record every agent-initiated retry as a new attempt. Keep scheduler automatic retry disabled unless the eval contract explicitly requires it. Ask before a material workspace, provider, credential, queue/resource, concurrency, or cost expansion; publication of durable baselines/catalog entries; physical robot movement; or any destructive deletion. Never delete a CloudML task or remote artifact without explicit authorization for the exact target and consequence.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 4d ago First seen · 328 lines · 99 tokens per session scan A 06bbc3508a60
cloudml-eval-ops is a skill published in the GitHub repository MiaoDX/roboclaws (6 stars, last pushed yesterday), licensed MIT. It adds 99 tokens to every session and 3,923 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 2026-08-31.
Other skills, from other repositories
remnote
Search, read, and write RemNote notes and personal knowledge base content via remnote-cli. Use for note-taking, journaling, tags, tables, and knowledge-base navigation; require confirm write before mutating commands.
remnote-kb-navigation
Template skill for navigating a user's RemNote knowledge base using root and top-level note IDs; customize before use.
smart-money-tracker
聪明钱追踪师 - 综合北向资金、龙虎榜、大宗交易、资金流向，追踪主力资金动向.
stock-analyst
股票技术分析专家 - 提供专业的 K 线形态、技术指标、资金流向与北向持仓分析.
kaggle
Unified Kaggle skill. Use when the user explicitly mentions Kaggle, kaggle.com, a Kaggle URL, Kaggle competitions, Kaggle datasets/models/notebooks, Kaggle forums/discussions/writeups, Kaggle benchmarks, hackathons hosted on Kaggle, Kaggle badges, or Kaggle account setup. Do not use for generic ML, GPU/TPU, notebook…
stock-screener
股票筛选器 - 按条件筛选全市场股票，找出符合策略的标的.