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 nebius/nebius-physical-ai --skill antioch-workbenchgit clone --depth 1 https://github.com/nebius/nebius-physical-aiWrote 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/nebius/nebius-physical-ai/antioch-workbench)<a href="https://agentmods.dev/skills/nebius/nebius-physical-ai/antioch-workbench"><img src="https://agentmods.dev/badge/skills/nebius/nebius-physical-ai/antioch-workbench/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/nebius/nebius-physical-ai/antioch-workbench"><img src="https://agentmods.dev/badge/skills/nebius/nebius-physical-ai/antioch-workbench.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 | $0.00046 | $0.01937 |
| Opus 5.5 | $0.00018 | $0.00775 |
| Sonnet 5.5 | $0.00009 | $0.00387 |
| Haiku 4.5 | $0.00005 | $0.00194 |
Grade A, and why
antioch-workbench 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 28d 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 — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Antioch and Workbench
Integration contract
Use the XR1 cookbook for the
executable procedure and pinned inputs. The operator module at
npa/src/npa/workflows/xr1_antioch/operator.py invokes the installed Antioch CLI
from an operator-owned Antioch project. Its commands are assets, runtime,
source, collect, fetch, publish, and compare. The antioch command
wraps native CLI arguments; exec launches attached remote argv through the
pinned SDK interpreter. Both resolve the NPA credential before launching.
The training stage is the workflow.xr1.finetune toolRef in
workflows/partners/antioch/xr1-antioch-finetune.yaml. It runs on Nebius and publishes
checkpoints to S3. Antioch collection, transfers, and closed-loop evaluation
are operator CLI/SDK steps around that YAML. Do not describe the entire loop as
one submitted workflow. The separate npa workbench antioch commands operate
the Workbench adapter and live OpenPI path; see the
partner catalog to choose a flow.
Authentication
- Inspect the installed CLI/SDK version and run the cookbook's
operator antioch --antioch-project <owned-directory> -- auth whoami --jsonprivately. This reports identity and credential source without printing the token, but organization, API, and user details still belong outside Git and public recordings. The tested XR1 adapter requiresantioch-sim==0.4.236. - Store a personal access token as
tokens.ANTIOCH_TOKENin~/.npa/credentials.yaml, honoringNPA_CONFIG_DIRwhen configured. Use the existing atomic 0600 credential writer. A secret-injectedANTIOCH_TOKENcan be saved throughnpa configure --save-env-credentials; this imports all supported credentials present in the environment and reports names only. Never request the token in chat or put its value in command arguments, shell history, committed files, workflow parameters, or recordings. - Resolve environment
ANTIOCH_TOKENbefore the NPA file value. With neither, retain the native CLI browser-login fallback (antioch auth login). Browser access and refresh credentials stay in Antioch's own auth store; never copy them into the personal-access-token field. When a personal access token is supplied, native login, logout, and organization switching refuse. Managed Antioch workspaces use their managed credentials first. If the identity is wrong, resolve the intended account before operating; do not silently switch organizations or remove credentials. - Use
npa/src/npa/clients/antioch.pyto supply the resolved credential only to local Antioch client subprocesses. The saved token is excluded from the generic token map, process-wide credential export, and shared worker env. Bare Antioch commands do not read the NPA store; use the operator wrapper when that stored token is required. Never add it to training YAML or S3. Verify access to the intended Antioch organization and project, an ownedantioch.yaml, a compatible engine, and available simulator capacity. A token authenticates the caller; it does not provision these prerequisites. - Configure Nebius and S3 separately through the selected NPA project. Follow
skills/atomic/health-preflight/SKILL.mdbefore provisioning or GPU submit, and the applicable model/software access procedure. An Antioch token does not grant Nebius access. Keep both providers' credentials on the operator; only the Nebius training worker needs the configured S3 credential delivery.
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.
- 28d ago Changed · +2 lines eff1c9903dd5
- 1mo ago First seen · 143 lines · 46 tokens per session scan A c7d2a3b3fcae
antioch-workbench is a skill published in the GitHub repository nebius/nebius-physical-ai (50 stars, last pushed today), licensed Apache-2.0. It adds 46 tokens to every session and 1,937 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-19.
Other skills, from other repositories
nextflow
Builds, runs, and debugs Nextflow DSL2 pipelines and nf-core workflows. Use for Nextflow, nf-core, .nf files, nextflow.config, processes/channels/operators, samplesheets, nf-test, modules/subworkflows, container and executor configuration, HPC/SLURM or cloud deployment, and failed or resumed pipeline runs.
latchbio-integration
Builds, registers, debugs, and operates bioinformatics workflows on Latch using the Python SDK, CLI, Latch Data and Registry, Nextflow, Snakemake, programmatic execution, and Latch MCP. Use when authoring or deploying Latch workflows, configuring resources or interfaces, moving data, integrating Registry, or launching…
modal-compute
Run explicitly chosen research benchmark or replication jobs on Modal's serverless infrastructure. Use when a Feynman research workflow needs burst remote GPU compute and the Modal CLI is available.
remote-compute-ssh
Evaluate and use SSH Remote Compute before choosing where to run GPU, high-memory, parallel, batch, model-inference, bioinformatics, or other long-running scientific work; supports short remote commands and asynchronous jobs with automatic harvest and analysis.
dnanexus-integration
DNAnexus cloud genomics platform. Build apps/applets, manage data (upload/download), dxpy Python SDK, run workflows, FASTQ/BAM/VCF, for genomics pipeline development and execution.
latchbio-integration
Latch platform for bioinformatics workflows. Build pipelines with Latch SDK, @workflow/@task decorators, deploy serverless workflows, LatchFile/LatchDir, Nextflow/Snakemake integration.