{"full_name":"mvschwarz/openrig","owner":"mvschwarz","name":"openrig","shard":"1","url":"https://github.com/mvschwarz/openrig","description":"Build your own network of agents from Claude Code, Codex and Pi: persistent teams with roles, shared context and owned work.","homepage":"https://openrig.dev","topics":["agent-harness","agent-orchestration","agent-skills","ai-coding","claude-code","cli","codex-cli","multi-agent","tmux","typescript","multi-agent-systems"],"tag_pages":["agent-harness","agent-orchestration","agent-skills","ai-coding","claude-code","cli","codex-cli","multi-agent","multi-agent-systems","tmux","typescript"],"language":"TypeScript","license":"Apache-2.0","avatar":"https://avatars.githubusercontent.com/u/171890339?u=0ce5231143a9bdeb8efa79c37490f3a8d03499df&v=4","created_at":"2026-04-01T06:22:40Z","pushed_at":"2026-11-15T19:30:51Z","metrics":{"stars":8861,"forks":759,"watchers":27,"issues":603,"prs":2641,"releases":66,"issues_open":163,"issues_closed":440,"prs_open":66,"prs_closed":155,"prs_merged":2420,"discussions":60},"growth":{"day":{"stars":12,"forks":6,"base_day":"2026-11-15"}},"history":{"dates":["2026-11-10","2026-11-11","2026-11-12","2026-11-13","2026-11-14","2026-11-15","2026-11-16"],"stars":[8798,8818,8832,8832,8843,8849,8861],"forks":[742,747,750,750,751,753,759],"watchers":[null,27,27,27,27,27,27],"issues_open":[null,156,153,153,155,156,163],"prs_open":[null,49,55,50,87,91,66]},"boards":{},"same_owner":[],"issues":[{"number":161,"title":"Daemon event loop saturates at scale: per-seat tmux polling (transcripts, structural, identity) spawns ~300 processes/s; transcript config leaks into seat env","comments":43,"created_at":"2026-09-29T19:37:21Z"},{"number":331,"title":"Stall checks miss unsubmitted wakes and flag waiting or working seats","comments":22,"created_at":"2026-10-01T16:40:10Z"},{"number":17,"title":"Codex hook-trust autoclear (OPR.0.4.3.33) does not suppress trust dialog on codex 0.155.1","comments":19,"created_at":"2026-09-23T18:45:57Z"},{"number":88,"title":"Windows Support","comments":17,"created_at":"2026-09-27T23:37:57Z"},{"number":2341,"title":"0.7.6: Claude seat `rig handover --source rebuild` always fails at start_agent (30 s readiness timeout) though the successor Claude is running","comments":16,"created_at":"2026-11-01T09:01:29Z"},{"number":81,"title":"Seats blocked on a permission prompt (needs_input) don't raise rig attention or any watchdog alert","comments":16,"created_at":"2026-09-27T20:14:35Z"},{"number":875,"title":"Keep seat projection out of the seat's cwd: deliver skills, hooks, MCP and guidance through launch flags","comments":15,"created_at":"2026-10-06T10:20:52Z"},{"number":313,"title":"Slack connector: support a DM or private channel for single-user setups","comments":15,"created_at":"2026-10-01T13:09:11Z"}],"discussions":[{"number":3171,"title":"The pi-runner pane names its seat and marks its own lines (follow-up to #2842)","comments":1,"category":"Ideas","created_at":"2026-11-13T13:13:06Z"},{"number":3117,"title":"An issue tracker as a network hub: give work to a team from GitHub Issues, Jira or Linear (asking before any PR)","comments":2,"category":"Ideas","created_at":"2026-11-13T01:28:48Z"},{"number":2974,"title":"Should a seat be able to own a git worktree? (optional, per-seat isolation)","comments":1,"category":"Ideas","created_at":"2026-11-11T09:13:49Z"},{"number":2919,"title":"Seats of the same role that learn together: pod-scoped skills grown from repeated procedures","comments":1,"category":"Ideas","created_at":"2026-11-10T23:22:21Z"},{"number":2873,"title":"Pipeworx as a default MCP next to Exa/Context7","comments":1,"category":"Ideas","created_at":"2026-11-10T01:39:48Z"},{"number":2862,"title":"Community runners: let the daemon talk back (cancel, wake, verified identity)","comments":4,"category":"Ideas","created_at":"2026-11-09T21:09:08Z"},{"number":2858,"title":"Attribute seat usage to the queue rows, slices and missions it was spent on","comments":3,"category":"Ideas","created_at":"2026-11-09T19:33:05Z"},{"number":2854,"title":"Community harness for nodecode (nodecode.ai) - built, in use, how to list it?","comments":4,"category":"Ideas","created_at":"2026-11-09T18:39:27Z"}],"latest_release":{"name":"OpenRig 0.7.16","tag":"v0.7.16","published_at":"2026-11-15T09:55:46Z","body":"## [0.7.16]\n\nOpenRig 0.7.16 is a patch for a person's first team and for recovering the same\nwork. A seat whose policy is the built-in `standard` or `open` keeps the team's\nlaunch default, so choosing `standard` no longer brings more prompts than\nchoosing nothing. The `rigs` skill tells a person's own agent to bring a seat's\npermission, trust and hook-review prompts to them as questions, then to run the\n`rig seat continue` a seat shows after its startup prompt is answered. A failed\nhandover's running successor can be adopted, a seat's runtime home can be\nrecorded without a relaunch, an upgrade keeps OpenRig's packaged recovery\nskills, and a send to a missing seat is kept and says so.\nThere are fixes across delivery, workflows and connected assistants, many\ncontributed by people using OpenRig.\n\n### Before you upgrade\n\n- **Stop the daemon, back up the database, install 0.7.16, then start the\n  daemon.** Your agents keep running through a daemon restart; don't use\n  `rig down`, which ends them.\n\n  ```sh\n  rig daemon stop\n  cp \"$(rig config get db.path)\" \"$(rig config get db.path).before-0.7.16\"\n  npm install -g @openrig/cli@0.7.16   # or your existing installation method\n  rig daemon start\n  ```\n\n- **Two migrations run before the daemon listens.** Migration 167 adds\n  `seat_handover_attempts` (for `rig seat handover --recover`), and migration\n  169 adds `seat_runtime_home_records` (for `rig seat record-runtime-home`). Going back to\n  0.7.15 after them isn't established; keep the backup. ([#3044](https://github.com/mvschwarz/openrig/pull/3044), [#3220](https://github.com/mvschwarz/openrig/pull/3220))\n- **The plugins:** the `openrig` plugin carries the CLI's version, 0.7.16.\n  openrig-core moves to 0.1.31, because its skills changed.\n\n### A person's first team\n\n- **A seat whose effective policy is the built-in `standard` or `open` keeps\n  the team launch default**, as a seat with no policy does. Any recorded policy\n  used to turn the team default off, so choosing `standard` brought more prompts\n  than choosing nothing: ordinary `rig` reads, git reads and tests all asked.\n  The member's policy decides over the rig's. An explicit seat permission\n  choice, a Codex envelope or named Codex profile, a Claude seat's own rules,\n  `none`, `locked`, custom, `auto` and `yolo`, and kernel, Pi and terminal seats\n  keep their meaning. This keeps the team default; it doesn't apply the whole\n  policy, and a seat that keeps it can still ask before the project's own\n  commands (see Known limits). The Claude team default includes ask rules for lifecycle commands\n  such as `rig up` and `rig down`. The Codex team default does not add\n  equivalent per-command ask rules; native settings and explicit permission\n  choices still apply. The launch plan now says whether a seat gets the team default's allowances,\n  instead of a bare `launch_posture=floor`. OpenRig can't see rules applied in\n  project or user native settings files; those seats keep the team default\n  alongside them, and keeping it doesn't loosen those rules. ([#3230](https://github.com/mvschwarz/openrig/pull/3230))\n- **The `rigs` skill tells a person's own agent to bring a seat's permission,\n  trust or hook-review prompt to them as a question** by default, instead of\n  answering it itself. It tells the agent to show the prompt, capture the seat\n  again before sending, and send the digit or letter the screen shows for their\n  choice, never their words, and to ask when the screen shows no key. A yes to\n  one prompt isn't a yes to the next, a \"don't ask again\" choice is picked only\n  when they picked it, and an instruction they gave outright lets the agent\n  decide that kind of prompt within what they said. The agent is told never to\n  record its own answer as the person's. The getting-started guide names each authentication, trust,\n  hook-review or permission prompt as the person's to answer before that seat\n  gets work. ([#3206](https://github.com/mvschwarz/openrig/pull/3206))\n- **After the pers","is_prerelease":false},"readme":true,"readme_zh":false,"news":[],"stale":false,"snapshot_at":"2026-11-16T09:23:14Z","history_start":"2026-09-06"}