{"id":"dcd7a37c-cbc3-4c95-bf5c-08c080eb10c3","slug":"paulclawx-browser-agent","kind":"skill","name":"browser-agent","summary":"[browser-agent] Never send a human to do a machine's job.","owner":"PaulClawX","repoUrl":"https://github.com/PaulClawX/browser-agent","license":"MIT","licenseTier":"permissive","trust":"community","stars":9,"match":null,"fewStars":false,"format":"skill","createdAt":"2026-05-08T18:49:51.000Z","addedAt":"2026-10-21T23:03:05.662Z","logoUrl":"https://avatars.githubusercontent.com/u/282689797?v=4","linkOnly":false,"collection":null,"skillPath":null,"categories":["browser-web"],"tags":["browser-agent","computer-use-agent","gui-agent"],"runtime":"python","local":true,"cloudCapable":false,"cloudCapability":"unknown","cloudReadyOnPlatform":false,"needsApiKey":false,"reviewed":false,"syncedAt":"2026-10-25T18:24:16.091Z","sample":false,"url":"https://conduits.dev/tools/paulclawx-browser-agent","language":"Python","transports":["stdio"],"install":{"pypi":"browser-agent"},"requirements":{"env":[],"upstreamMayCharge":false},"tools":null,"readme":"# Browser Agent\n\n<p align=\"center\">\n  <img src=\"docs/images/cover.png\" alt=\"Browser Agent cover\" width=\"70%\">\n</p>\n\n<p align=\"center\">\n  <strong>CDP-native browser automation runtime for agents, with editable skills and reliable Gemini image workflows.</strong>\n</p>\n\n<p align=\"center\">\n  • Real Chrome Control • Skill-Driven Automation • Upload-Verified Image Generation •\n</p>\n\n<p align=\"center\">\n  <a href=\"#-features\">Features</a> •\n  <a href=\"#-quick-start\">Quick Start</a> •\n  <a href=\"#-workflows\">Workflows</a> •\n  <a href=\"#-safety-model\">Safety Model</a> •\n  <a href=\"SKILL.md\">Agent Guide</a>\n</p>\n\n<p align=\"center\">\n  <a href=\"LICENSE\"><img alt=\"License\" src=\"https://img.shields.io/badge/license-MIT-green\"></a>\n  <img alt=\"Status\" src=\"https://img.shields.io/badge/status-0.1.0--alpha-orange\">\n  <img alt=\"Python\" src=\"https://img.shields.io/badge/Python-3.9%2B-blue\">\n  <img alt=\"CDP\" src=\"https://img.shields.io/badge/CDP-native-111827\">\n  <img alt=\"Codex\" src=\"https://img.shields.io/badge/Codex-skill--ready-111827\">\n</p>\n\n## Quick Navigation\n\n> [!TIP]\n> **I'm a human** -> Read this README for install, setup, and safe workflows.\n>\n> **I'm an agent** -> Read [SKILL.md](SKILL.md) for operation rules and execution patterns. (Recommended)\n\n`browser-agent` is a minimal runtime that lets agents control your real Chrome session directly over CDP, while keeping task logic editable in-repo.\n\n- **For operators**: one command surface for browser actions, diagnostics, and updates.\n- **For agents**: stable helper APIs (`new_tab`, `js`, `click_at_xy`, `upload_file`, raw `cdp`).\n- **For reliability**: interaction skills and domain skills to encode repeatable mechanics.\n\n## Quick Start\n\n### For Agent (Recommended)\nTell your coding agent:\n```text\nInstall Browser Agent from https://github.com/PaulClawX/browser-agent and set it up to control my Chrome\n```\n\n### For Human\n```bash\ngit clone https://github.com/PaulClawX/browser-agent && cd browser-agent && uv tool install -e . && browser-agent --doctor\n```\n\n## Browser Connection\n\n### Attach to your normal Chrome profile\n\n1. Open `chrome://inspect/#remote-debugging`\n2. Enable `Allow remote debugging for this browser instance`\n3. Accept the Chrome allow popup when prompted\n\n![Remote Debugging Setup](docs/images/setup-remote-debugging.png)\n\n4. Re-run:\n```bash\nbrowser-agent -c 'print(page_info())'\n```\n\n> See [install.md](install.md) for full setup.\n\n\n\n## Workflows\n\n| Tier | Workflow | Expected Behavior |\n|---|---|---|\n| Stable | General browser automation | Deterministic tab + DOM + input operations through CDP helpers |\n| Stable | Upload-driven tasks | Upload confirmation before submit; fail-fast if upload isn't verifiable |\n| Stable | Gemini image generation/editing | Prompt + reference flow with strict upload-first gating and export |\n| Stable | Diagnostics and lifecycle | `--doctor`, daemon auto-start, update checks |\n| Best-effort | Complex anti-bot sites | Fallback to coordinate actions, retries, and skill-specific patterns |\n\n\n## Project Layout\n\n- `src/browser_harness/` - core runtime modules\n- `SKILL.md` - operator rules for day-to-day use\n- `install.md` - first-time install and connection\n- `docs/interaction-skills/` - reusable browser mechanics playbooks\n- `src/agent-workspace/agent_helpers.py` - task-specific helper extensions\n- `docs/domain-skills/` - site-specific playbooks\n\n\n\n\n## Core Contributors and Maintainers\n\n<table>\n  <tr>\n    <td align=\"center\">\n      <a href=\"https://github.com/paulpanwang\">\n        <img src=\"https://images.weserv.nl/?url=https://paulpanwang.github.io/images/paul.jpg&h=100&w=100&fit=cover&mask=circle&maxage=7d\" width=\"100px;\" alt=\"Panwang Pan\"/>\n      </a>\n      <br />\n      <sub><b>Panwang Pan</b></sub>\n      <br />\n      <sub><a href=\"mailto:paulpanwang@gmail.com\">paulpanwang@gmail.com</a></sub>\n    </td>\n    <td align=\"center\">\n      <a href=\"https://angericky.github.io/\">\n        <img src=\"https://images.weserv.nl/?url=https://i.loli.net/2019/01/07/5c336be8ba185.jpg&h=100&w=100&fit=cover&mask=circle&maxage=7d\" width=\"100px;\" alt=\"Jingjing Zhao\"/>\n      </a>\n      <br />\n      <sub><b>Jingjing Zhao</b></sub>\n      <br />\n      <sub><a href=\"mailto:jingjingbudlet@gmail.com\">jingjingbudlet@gmail.com</a></sub>\n    </td>\n  </tr>\n</table>\n\n\n## 📧 Contact\n\nFeel free to open an issue if you have any questions or suggestions. If this project helps you, please give it a ⭐ Star!\n\n## Acknowledgements\n\nThis project builds on and is inspired by the following open-source work:\n\n- [browser-use/browser-harness](https://github.com/browser-use/browser-harness) - the primary code and architecture source.\n- [OpenClaudex/openreview-agent](https://github.com/OpenClaudex/openreview-agent) - OpenReview dry-run workflow inspiration.\n"}