Skip to content
Back to search

yolodex

agent skills for autonomous data labeling, winner at openai codex hackathon 2026

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/qtzx06--yolodex/badge.svg)](/listings/qtzx06--yolodex)

HTML

<a href="/listings/qtzx06--yolodex"><img src="/api/listings/qtzx06--yolodex/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

git clone https://github.com/qtzx06/yolodex && cd yolodex
bash setup.sh

requirements: - macos, linux, or windows - python 3.11+ - [codex cli](https://github.com/openai/codex) (latest — supports forked skill contexts and per-skill model overrides)

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 189 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. <h1 align="center">yolodex</h1>
  2. <div align="center">
  3. ![GitHub last commit](https://img.shields.io/github/last-commit/qtzx06/yolodex?style=for-the-badge&labelColor=ffffff&color=8ecaff)
  4. ![GitHub Repo stars](https://img.shields.io/github/stars/qtzx06/yolodex?style=for-the-badge&labelColor=ffffff&color=8ecaff)
  5. ![GitHub repo size](https://img.shields.io/github/repo-size/qtzx06/yolodex?style=for-the-badge&labelColor=ffffff&color=8ecaff)
  6. ![License: MIT](https://img.shields.io/badge/license-MIT-8ecaff?style=for-the-badge&labelColor=ffffff)
  7. ![PRs welcome](https://img.shields.io/badge/PRs-welcome-8ecaff?style=for-the-badge&labelColor=ffffff)
  8. </div>
  9. <div align="center">
  10. codex-native yolo dataset + training agent skills for any unlabeled data. codex skills help determine how to properly label and splice your data.
  11. </div>
  12. [![star history chart](https://api.star-history.com/svg?repos=qtzx06/yolodex&type=Date)](https://star-history.com/#qtzx06/yolodex&Date)
  13. ## architecture
  14. ```mermaid
  15. flowchart LR
  16. A["data source (any unlabeled data)"] --> B["collect"]
  17. B --> C["label"]
  18. C --> D["augment"]
  19. D --> E["train"]
  20. E --> F["eval"]
  21. C -. "parallel subagents in git worktrees" .-> C
  22. F --> G{"meets target?"}
  23. G -- "no" --> C
  24. G -- "yes" --> H["complete"]
  25. ```
  26. ## components
  27. - skills runtime: [codex skills](https://developers.openai.com/codex/skills/)
  28. - pipeline skills: `collect`, `label`, `augment`, `train`, `eval`, `yolodex`
  29. - orchestration loop: `yolodex.sh` + `AGENTS.md`
  30. - shared helpers: `shared/utils.py`
  31. ## installation
  32. ```bash
  33. git clone https://github.com/qtzx06/yolodex && cd yolodex
  34. bash setup.sh
  35. ```
  36. requirements:
  37. - macos, linux, or windows
  38. - python 3.11+
  39. - [codex cli](https://github.com/openai/codex) (latest — supports forked skill contexts and per-skill model overrides)
  40. ## codex workflow (main path)
  41. 1. start codex in repo root
  42. 2. use the `yolodex` skill to gather config
  43. 3. run labeling with subagents when needed
  44. 4. iterate until eval target is met
  45. example:
  46. ```text
  47. $ codex
  48. > use the yolodex skill to train a detector from this video: https://youtube.com/...
  49. > classes: player, weapon, vehicle
  50. ```
  51. codex determines how many subagents to spawn for labeling based on the dataset size:
  52. ```bash
  53. bash .agents/skills/label/scripts/dispatch.sh <n>
  54. ```
  55. ## run modes
  56. autonomous loop:
  57. ```bash
  58. bash yolodex.sh
  59. ```
  60. manual skills:
  61. ```bash
  62. uv run .agents/skills/collect/scripts/run.py
  63. bash .agents/skills/label/scripts/dispatch.sh 4
  64. uv run .agents/skills/augment/scripts/run.py
  65. uv run .agents/skills/train/scripts/run.py
  66. uv run .agents/skills/eval/scripts/run.py
  67. ```
  68. ## config
  69. minimal `config.json`:
  70. ```json
  71. {
  72. "project": "my-project",
  73. "video_url": "https://youtube.com/watch?v=YOUR_VIDEO",
  74. "classes": ["player", "weapon", "vehicle"],
  75. "label_mode": "codex",
  76. "target_accuracy": 0.75,
  77. "num_agents": 4,
  78. "fps": 1,
  79. "yolo_model": "yolov8n.pt",
  80. "epochs": 50,
  81. "train_split": 0.8,
  82. "seed": 42
  83. }
  84. ```
  85. key fields:
  86. | field | default | description |
  87. |---|---|---|
  88. | `project` | `""` | output namespace under `runs/<project>/` |
  89. | `video_url` | `""` | youtube url or local video path |
  90. | `classes` | `[]` | target classes |
  91. | `label_mode` | `"codex"` | label strategy |
  92. | `target_accuracy` | `0.75` | mAP@50 stop threshold |
  93. | `num_agents` | `4` | parallel label workers |
  94. | `fps` | `1` | extraction fps |
  95. | `yolo_model` | `"yolov8n.pt"` | base yolo checkpoint |
  96. | `epochs` | `50` | train epochs |
  97. | `train_split` | `0.8` | train/val split |
  98. | `seed` | `42` | deterministic split seed |
  99. ## outputs
  100. - eval: `runs/<project>/eval_results.json`
  101. - frames: `runs/<project>/frames/`
  102. - label previews: `runs/<project>/frames/preview/`
  103. - trained weights: `runs/<project>/weights/best.pt`
  104. ## repo layout
  105. ```text
  106. yolodex/
  107. ├── .agents/skills/
  108. │ ├── yolodex/
  109. │ ├── collect/
  110. │ ├── label/
  111. │ ├── augment/
  112. │ ├── train/
  113. │ └── eval/
  114. ├── shared/utils.py
  115. ├── pipeline/main.py
  116. ├── docs/
  117. ├── yolodex.sh
  118. ├── setup.sh
  119. ├── AGENTS.md
  120. ├── config.json
  121. ├── progress.txt
  122. └── pyproject.toml
  123. ```
  124. ## docs
  125. - [docs/usage.md](docs/usage.md)
  126. - [docs/models.md](docs/models.md)
  127. - [docs/skills.md](docs/skills.md)
  128. - [docs/architecture.md](docs/architecture.md)
  129. - [docs/changelog.md](docs/changelog.md)
  130. ## contributing
  131. contributions are welcome.
  132. standard flow:
  133. 1. fork the repo and create a branch from `main`
  134. 2. keep changes scoped and explain the why in your pr
  135. 3. run relevant checks before opening the pr
  136. 4. open a pull request with:
  137. - clear summary
  138. - test/validation notes
  139. - screenshots/log snippets when relevant
  140. if you’re planning a bigger change, open an issue first so we can align on direction.
  141. ## license
  142. mit
  143. ---
  144. <p align="center">made with love at openai ♡</p>

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.