# Cvat Pull

> Pull human-corrected annotations from CVAT into local YOLO dataset for training.

- **Type:** Skill
- **Install:** `agentstack add skill-macroman5-autotrain-yolo-cvat-pull`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [MacroMan5](https://agentstack.voostack.com/s/macroman5)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [MacroMan5](https://github.com/MacroMan5)
- **Source:** https://github.com/MacroMan5/autotrain-yolo/tree/master/.claude/skills/cvat-pull

## Install

```sh
agentstack add skill-macroman5-autotrain-yolo-cvat-pull
```

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

## About

# Pull CVAT Annotations

Pull human-corrected annotations from CVAT into your local dataset.

## Pre-Flight Checklist
- [ ] `yolo-project.yaml` has `cvat:` section with url and project_id
- [ ] `CVAT_ACCESS_TOKEN` env var is set
- [ ] CVAT instance is reachable (default: http://localhost:8080)

## Workflow

### 1. Identify What to Pull
- Ask the user which CVAT task or project to pull
- Or read the default `project_id` from `yolo-project.yaml`

### 2. Pull Annotations
```bash
yolo-cvat pull --task 
# or
yolo-cvat pull --project 
```

### 3. Validate the Downloaded Dataset
```bash
yolo-validate 
```
Check for:
- Valid data.yaml with correct class names
- Image/label count matches expectations
- No annotation errors

### 4. Compare with Existing Dataset
If the user already has a local dataset:
- Compare class distributions
- Check for new images vs corrections
- Suggest merge strategy if combining

### 5. Merge if Needed
```bash
yolo-merge --sources   --output 
```

## Decision Tree

```
Has existing local dataset?
├── Yes → Compare distributions → Suggest yolo-merge
└── No → Set as primary dataset

Validation warnings found?
├── Yes → Flag issues, ask before training
└── No → Ready to train

Class distribution changed?
├── Significantly → Warn user, may affect model balance
└── Minor → Proceed normally
```

## Guardrails
- NEVER overwrite an existing dataset directory without user confirmation
- ALWAYS run yolo-validate after pulling
- Report class distribution changes clearly
- If merge is needed, show the user what will change before executing

## Source & license

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

- **Author:** [MacroMan5](https://github.com/MacroMan5)
- **Source:** [MacroMan5/autotrain-yolo](https://github.com/MacroMan5/autotrain-yolo)
- **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-macroman5-autotrain-yolo-cvat-pull
- Seller: https://agentstack.voostack.com/s/macroman5
- 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%.