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 skills add metaspartan/cybara --skill huggingface-jobsgit clone --depth 1 https://github.com/metaspartan/cybaraWrote 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/metaspartan/cybara/huggingface-jobs)<a href="https://agentmods.dev/skills/metaspartan/cybara/huggingface-jobs"><img src="https://agentmods.dev/badge/skills/metaspartan/cybara/huggingface-jobs/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/metaspartan/cybara/huggingface-jobs"><img src="https://agentmods.dev/badge/skills/metaspartan/cybara/huggingface-jobs.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00036 | $0.00429 |
| Opus 5 | $0.00018 | $0.00215 |
| Sonnet 5 | $0.00007 | $0.00086 |
| Haiku 4.5 | $0.00004 | $0.00043 |
Grade A, and why
huggingface-jobs 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 11d 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.
What it actually says
Hugging Face Jobs
Use hf jobs for remote workloads. Jobs consume paid compute and run outside the local workspace.
Preflight
- Run
hf version,hf jobs --help, andhf auth whoami. - Verify the account or organization has access and sufficient credits.
- Inspect available hardware with
hf jobs hardware. - Define the command or UV script, dependencies, hardware flavor, namespace, timeout, persistence path, and expected cost boundary.
- Present the exact launch plan and obtain confirmation before starting paid compute.
Never transmit a token in source, command arguments, or logs. Use the authenticated CLI session and pass sensitive values through the Jobs secrets mechanism.
Execution
Prefer a self-contained UV script for Python workloads:
hf jobs uv run --detach --flavor gpu-t4-small --timeout 2h train.py
Use a container only when the workload requires a specific system image:
hf jobs run --detach --flavor cpu-basic --timeout 30m python:3.12 python -c "print('ready')"
Start with the cheapest representative smoke run. Do not scale to the full dataset or larger hardware until inputs, imports, persistence, and metrics are verified.
Monitoring
hf jobs ps --all --format json
hf jobs inspect <job-id>
hf jobs logs <job-id> --tail 200
hf jobs stats <job-id>
hf jobs cancel <job-id>
After launch, report the job ID, URL, namespace, hardware, timeout, and where outputs will persist. Do not poll indefinitely. Check status when the user asks or when an active task explicitly requires completion.
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.
- 11d ago First seen · 48 lines · 36 tokens per session scan A 06dbe721d57a
huggingface-jobs is a skill published in the GitHub repository metaspartan/cybara (28 stars, last pushed today), licensed MIT. It adds 36 tokens to every session and 429 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
vLLM
Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.
ecc-cost-aware-llm-pipeline
Cost optimization patterns for LLM API usage — model routing by task complexity, budget tracking, retry logic, and prompt caching. Use when LLM spend needs to come down, or when routing tasks across model tiers and budgets.
serving-llms-vllm
Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.
serving-llms-vllm
Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.
vercel-deploy
Deploy applications and websites to Vercel. Use this skill when the user requests deployment actions such as "Deploy my app", "Deploy this to production", "Create a preview deployment", "Deploy and give me the link", or "Push this live". No authentication required - returns preview URL and claimable deployment link.
bigquery-ai-ml
Skill for BigQuery AI and Machine Learning queries using standard SQL and AI. functions (preferred over dedicated tools).