{"id":47777175,"url":"https://github.com/opencsgs/csgclaw","last_synced_at":"2026-05-24T11:01:32.235Z","repository":{"id":347483890,"uuid":"1191181340","full_name":"OpenCSGs/csgclaw","owner":"OpenCSGs","description":"Your own personal AI team.","archived":false,"fork":false,"pushed_at":"2026-04-15T02:45:03.000Z","size":5394,"stargazers_count":0,"open_issues_count":0,"forks_count":1,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-04-15T04:24:04.920Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Go","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/OpenCSGs.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":"AGENTS.md","dco":null,"cla":null}},"created_at":"2026-03-25T01:56:39.000Z","updated_at":"2026-04-15T02:45:06.000Z","dependencies_parsed_at":null,"dependency_job_id":"502a5bd4-d923-441c-81ad-22a12c067125","html_url":"https://github.com/OpenCSGs/csgclaw","commit_stats":null,"previous_names":["opencsgs/csgclaw"],"tags_count":7,"template":false,"template_full_name":null,"purl":"pkg:github/OpenCSGs/csgclaw","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OpenCSGs%2Fcsgclaw","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OpenCSGs%2Fcsgclaw/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OpenCSGs%2Fcsgclaw/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OpenCSGs%2Fcsgclaw/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/OpenCSGs","download_url":"https://codeload.github.com/OpenCSGs/csgclaw/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OpenCSGs%2Fcsgclaw/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":31869050,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-15T15:24:51.572Z","status":"online","status_checked_at":"2026-04-16T02:00:06.042Z","response_time":69,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-04-03T12:04:46.370Z","updated_at":"2026-05-24T11:01:32.225Z","avatar_url":"https://github.com/OpenCSGs.png","language":"Go","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cp align=\"center\"\u003e\n  \u003cimg src=\"assets/logo.png\" alt=\"CSGClaw logo\" width=\"600\" /\u003e\n\u003c/p\u003e\n\n\u003cp align=\"center\"\u003e\n  English | \u003ca href=\"./README.zh.md\"\u003e中文\u003c/a\u003e\n\u003c/p\u003e\n\n# CSGClaw\n\n\u003e Your Personal AI Team\n\nCSGClaw is a multi-agent collaboration platform built by OpenCSG — designed around one practical question: **once work becomes non-trivial, how do you get a group of AI agents to operate like a team, without the system becoming heavy or painful to set up?**\n\n## Install\n\n**macOS / Linux:**\n\n```bash\ncurl -fsSL https://csgclaw.opencsg.com/install.sh | bash\n```\n\n**Windows (PowerShell):**\n\n```powershell\ncurl.exe -fsSL https://csgclaw.opencsg.com/install.ps1 | powershell -ExecutionPolicy Bypass -Command -\n```\n\nThe installers download a prebuilt release bundle, install it into user-local directories, and put `csgclaw` on your `PATH`.\n\n- `install.sh` currently supports macOS arm64, Linux amd64, and Linux arm64.\n- `install.ps1` currently supports Windows amd64.\n\nOfficial release bundles are also published for macOS amd64 and Windows amd64. Windows bundles currently default to Docker and require a working local Docker installation.\n\n**Build from source:**\n\n```bash\nmake build\n```\n\nFor most users, the install script above is the simpler option.\n\n## Quick Start\n\n```bash\ncsgclaw serve\n```\n\nCSGClaw will open the IM workspace in your browser automatically when possible. If it does not, open the printed URL manually (for example `http://127.0.0.1:18080/`).\n\n## Configuration\n\n`csgclaw serve` uses a local config with server, bootstrap, sandbox, and channel settings, and auto-creates any missing bootstrap state on first run. Agent model/provider profiles are stored in agent state and managed from the Web UI. See [docs/config.md](docs/config.md) for sandbox provider options, Worker override examples, and agent profile details.\n\nMaintainer-facing architecture, API, and IM thread design notes live in [docs/architecture.md](docs/architecture.md), [docs/api.md](docs/api.md), and [docs/im-threads.md](docs/im-threads.md).\n\n## Features\n\n- **Multi-agent coordination** — work with a team of specialized agents through a single coordination point, not a pile of chat windows\n- **One-click install** — prebuilt releases for macOS arm64/amd64, Linux amd64/arm64, and Windows amd64; up and running in minutes\n- **WebUI out of the box** — browser-based workspace available immediately after `csgclaw serve`\n- **Multi-channel support** — connect Feishu, WeChat, Matrix, or other channels when needed\n- **Isolated execution** — each Worker runs in a secure sandbox with security boundaries enabled by default\n- **Role-based Workers** — specialize Workers for frontend, backend, testing, docs, research, and more\n\n## What CSGClaw Is\n\nCSGClaw gives you one **Manager** and a set of specialized **Workers**, so instead of juggling isolated agents, you work through a single coordination point for defining goals, breaking down work, assigning roles, tracking progress, and collecting results.\n\n```text\n┌────────────────────────────────────────────────────────────┐\n│                         CSGClaw                            │\n│  ┌──────────────────────────────────────────────────────┐  │\n│  │ Manager — understands goals, plans, coordinates      │  │\n│  └──────────────────────────────────────────────────────┘  │\n│               ↓                      ↓                     │\n│        Worker Alice            Worker Bob                  │\n│          frontend                 backend                  │\n│                                                            │\n│   WebUI / Feishu / WeChat / Matrix / other channels        │\n└────────────────────────────────────────────────────────────┘\n                      ↑ you make decisions\n```\n\n**Manager** — receives your goals, decomposes tasks, selects Workers, tracks progress, and consolidates results.\n\n**Workers** — role-specific executors (frontend, backend, testing, docs, research…). Specialization keeps context clean and reduces role confusion.\n\n**Sandbox** — Worker execution is isolated by the configured sandbox provider. Official bundles default to **BoxLite** when they include a bundled `boxlite`, and default to **Docker** otherwise.\n\n**Interface** — WebUI out of the box; Feishu, WeChat, Matrix, and other channels available as integrations.\n\n## A Typical Workflow\n\n```text\nYou: Build a web app prototype — landing page, login, and basic admin view.\n\nManager: Splitting into tasks.\n  · Alice → landing page \u0026 login UI\n  · Bob   → backend APIs \u0026 data model\n  · Carol → integration checks\n\nYou: Add GitHub login to the login flow.\n\nManager: Updating Alice and Bob.\n\nCarol: Login response is missing the user avatar field.\n\nManager: Bob updates the API first; Alice updates the UI once the field contract is confirmed.\n```\n\nThe key isn't that multiple agents exist — it's that **their collaboration is organized**.\n\n## Design Principles\n\n**PicoClaw by default, extensible by design.**\nCSGClaw uses PicoClaw as its lightweight default Agent Runtime, keeping the Manager fast to start and cheap to run. The runtime remains pluggable, so deployments can integrate alternatives such as OpenClaw when needed.\n\n**BoxLite by default, sandbox-agnostic by design.**\nIsolation is non-negotiable. BoxLite is the preferred default because it is lightweight, fast to start, and convenient for local-first workflows. When BoxLite is unavailable, Docker is fully supported as a mature cross-platform fallback, and teams can switch sandbox providers explicitly when their environment requires it.\n\n**WebUI first, channel-agnostic by design.**\nMany multi-agent systems are tightly coupled to one messaging protocol. CSGClaw ships with a built-in WebUI so you can start immediately, while keeping other channels (Feishu, WeChat, Matrix) as optional integrations — not assumptions.\n\n## Who It Is For\n\n- Independent developers who want an AI team, not just a single assistant\n- Small teams that want lower-friction multi-agent collaboration\n- Users who value fast startup, lighter runtime, and sensible defaults\n\n## Acknowledgement\n\nCSGClaw is informed by ideas explored in HiClaw around multi-agent usability, while placing stronger emphasis on lightweight runtime, easier local startup, and a platform model not bound to a single communication channel.\n\n## License\n\nCSGClaw is licensed under the Apache License 2.0. See [LICENSE](LICENSE) for details.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fopencsgs%2Fcsgclaw","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fopencsgs%2Fcsgclaw","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fopencsgs%2Fcsgclaw/lists"}