# Liab Deploy

> >

- **Type:** Skill
- **Install:** `agentstack add skill-neuroanalytics-data-science-harness-liab-deploy`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [neuroanalytics](https://agentstack.voostack.com/s/neuroanalytics)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [neuroanalytics](https://github.com/neuroanalytics)
- **Source:** https://github.com/neuroanalytics/data-science-harness/tree/main/plugins/disseminate/skills/liab-deploy

## Install

```sh
agentstack add skill-neuroanalytics-data-science-harness-liab-deploy
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Skill: liab-deploy

Give the project a **data-sovereign** distribution channel: self-hosted infrastructure that serves
the DataLad dataset over git-annex remotes, so the lab controls where its data lives while keeping
the same Distributability a cloud sibling provides. The infra is declarative (pyinfra), so the
deployment itself is reproducible. You delegate sibling registration and the save to the **datalad
doer**.

Load `plugins/disseminate/references/liab-deployments.md` for the deployment layout and mapping
before generating.

## When to use
- The project wants to self-host its dataset (data sovereignty, institutional policy), alongside or
  instead of a cloud archive.
- Do NOT use to push to an existing sibling (`disseminate/publish`) or to mint a DOI
  (`dataset-release`) — this scaffolds the *destination infrastructure*.

## Steps
1. **Gather targets** — the host(s) that will run the deployment and whether Forgejo (git host) +
   git-annex serving are both wanted. Confirm the operator has access to the target hosts.
2. **Scaffold the deployment** (per the reference) at `liab/`: `inventory.py` (hosts), `deploy.py`
   (pyinfra operations for Forgejo + git-annex special remote), and `config/`. Do not run the
   deployment for the user — scaffold it and show the `pyinfra inventory.py deploy.py` command.
3. **Register the sibling (datalad doer)** — once the store is up, delegate:
   > "siblings: register the self-hosted Forgejo/git-annex store as a sibling (`create-sibling` /
   > the annex special remote), with a storage `--publish-depends` so annexed content is served."
   Then the user pushes with `disseminate/publish`.
4. **Register + log** — record the deployment path (and, once live, the sibling name) under a product
   (kind `other`) `outputs[]`; append
   `{ ts, op: liab-deploy, stage: disseminate, note: "Lab-in-a-Box deploy scaffold; sibling ", branch:  }`.
5. **Save** — delegate to the datalad doer: "save: `datalad save -m 'liab-deploy: scaffold self-hosted serving'`."
6. **Report** — the deployment path, how to run it, the registered sibling (once live), and the next
   step: `disseminate/publish` to push to the self-hosted store, then `link-outputs` to relate the
   mirror to the dataset (`IsVariantFormOf`).

## Constraints
- Scaffold, do not deploy: never run `pyinfra` or provision remote hosts on the user's behalf —
  present the command and let the operator run it against their own infrastructure.
- The self-hosted store is a *destination* — it reuses the standard sibling/publish/provenance flow
  (register via the datalad doer, push via `publish`), it does not bypass it.
- Ensure annexed content is actually served (a storage `--publish-depends`) so a clone can
  `datalad get` from the self-hosted store — otherwise it distributes history without data.
- Record under a product's `outputs[]`; keep `log:` append-only and the ledger schema-valid.
  Delegate sibling/save operations to the datalad doer.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [neuroanalytics](https://github.com/neuroanalytics)
- **Source:** [neuroanalytics/data-science-harness](https://github.com/neuroanalytics/data-science-harness)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-neuroanalytics-data-science-harness-liab-deploy
- Seller: https://agentstack.voostack.com/s/neuroanalytics
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.