# Qc Review

> >

- **Type:** Skill
- **Install:** `agentstack add skill-neuroanalytics-data-science-harness-qc-review`
- **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/govern/skills/qc-review

## Install

```sh
agentstack add skill-neuroanalytics-data-science-harness-qc-review
```

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

## About

# Skill: qc-review

Assess whether the dataset is in good shape to build on and release: is it **BIDS-valid**
(Self-containment), and how does the project score across STAMPED? This is a read-only review that
turns the current state into a short report with concrete next actions. You delegate validation to
the **bids doer** and state inspection to the **datalad doer**; you do not modify anything.

## When to use
- Before grouping/releasing a product, at the end of curation, or whenever the user wants a quality
  or reproducibility check.
- Do NOT use to fix issues — this skill only diagnoses. It routes each gap to the skill that fixes
  it (`curate/annotate`, `curate/raw-to-bids`, `analyze/checkpoint`, `disseminate/publish`, …).

## Steps
1. **BIDS validation (bids doer)** — delegate:
   > "validate this dataset against BIDS and report result (valid/invalid/unverified), error and
   > warning counts, and any dataset_description/participants/sidecar gaps."
   Relay the summary. If the validator is absent, report `unverified` with the install hint — never
   claim validity that was not checked.
2. **State inspection (datalad doer)** — delegate a read-only check:
   > "status + log: is the working tree clean, what is the current branch, is a sibling configured?"
3. **STAMPED self-assessment** — from the bids result, the datalad state, and the ledger
   (`project.yaml`), score each letter and name the gap-closing skill:
   - **S** self-contained → BIDS valid + `dataset_description.json`/`README` present (`curate/*`)
   - **T** tracked → working tree clean, runs recorded (`analyze/checkpoint`, `run-comparison`)
   - **A** actionable → analyses are provenanced `datalad run`s, not ad-hoc shell
   - **M** metadata → data dictionary / sidecars / controlled terms present (`curate/annotate`)
   - **P/E** portable/ephemeral → a container recipe exists and analyses run in it
     (`process/run-pipeline`, `run-comparison`)
   - **D** distributable → a sibling exists / a product is released (`disseminate/publish`,
     `dataset-release`)
4. **Log it** — append `{ ts, op: qc-review, stage: qc, note: "bids ; STAMPED gaps: ", branch:  }`, then delegate the save to the datalad doer.
5. **Report** — a compact scorecard: BIDS result (errors/warnings), the STAMPED letters that are
   satisfied vs. the gaps, and for each gap the single skill that closes it. Recommend the highest
   priority next action.

## Constraints
- Read-only: never modify, rename, or "fix" dataset files here — diagnose and route to the fixing
  skill. Validation and state inspection go through the bids and datalad doers respectively.
- Never assert BIDS validity the bids doer did not verify (an absent validator is `unverified`, not
  a pass).
- The STAMPED assessment is a snapshot, not a gate — report it plainly with gaps; do not block other
  work. Keep `log:` append-only and the ledger schema-valid.

## 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-qc-review
- 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%.