# Annotate

> Claude visually inspects images and corrects/creates YOLO bounding box annotations using multimodal vision.

- **Type:** Skill
- **Install:** `agentstack add skill-macroman5-autotrain-yolo-annotate`
- **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/annotate

## Install

```sh
agentstack add skill-macroman5-autotrain-yolo-annotate
```

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

## About

# Annotate

Claude vision annotation correction using multi-pass refinement.

**STATUS: EXPERIMENTAL**

## Approach
1. YOLO inference at low confidence (0.15) → precise boxes
2. Claude vision review → keeps good, removes bad, identifies missed
3. For missed objects → YOLO on cropped region
4. Final Claude validation

## Workflow
1. Read `yolo-project.yaml` for class names
2. For each image:
   a. Draw existing annotations: `python scripts/draw_annotations.py`
   b. Claude reads annotated image
   c. Evaluate: keep / correct / remove / add
   d. Write corrected YOLO label

## Coordinate Guide
- (0.0, 0.0) = top-left, (1.0, 1.0) = bottom-right
- (0.5, 0.5) = center
- Width/height relative to image

## Limitations
- ~10-15% coordinate error
- Best for objects > 5% of image area
- Prefer correcting existing boxes over creating from scratch

## 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-annotate
- 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%.