factory-rsi reference
A seven-agent team that plans, builds, tests, reviews, tries out and releases changes, and waits for you before anything is published. It's kept as a reference, with no verified run on current OpenRig.
Listed by OpenRig maintainers · Reviewed for listing on 6 Oct 2026 at commit f1d1d8fa4beb · not a security audit
github.com/mvschwarz/openrig-world · rigs/launch/factory-rsi · commit f1d1d8fa4beb
The team
How much OpenRig has tested it
For the recommended configuration, from OpenRig's published test status.
- UntestedOpenRig ran this team before but hasn't tested it recently; it's here as a reference structure.
Harness checks
Whether each harness installs and starts. Separate from the team's test status above.
- Claude Code on macOSPassedClaude Code 2.1.290 · OpenRig 0.6.6 · 6 Oct 2026
- Claude Code on LinuxPassedClaude Code 2.1.299 (updated to 2.1.302) · OpenRig 0.6.11 · 18 Oct 2026
- Codex on macOSPassedCodex 0.160.0 · OpenRig 0.6.6 · 6 Oct 2026
- Codex on LinuxPassedCodex 0.163.0 · OpenRig 0.6.11 · 18 Oct 2026
What it will do
Which agents start
- plan.planner on claude-code, profile default, working folder "."
- build.implementer on claude-code, profile default, working folder "."
- check.qa on codex, profile default, working folder "."
- review.reviewer on codex, profile default, working folder "."
- dogfood.tester on codex, profile default, working folder "."
- release.manager on claude-code, profile default, working folder "."
- orch.lead on claude-code, profile default, working folder "."
What they can do
- plan.planner: Can run shell commands as the launching user, subject to runtime and host policy. acceptEdits (conditional launch floor). OpenRig's default.
- build.implementer: Can run shell commands as the launching user, subject to runtime and host policy. acceptEdits (conditional launch floor). OpenRig's default.
- check.qa: Can run shell commands as the launching user, subject to runtime and host policy. workspace-write (conditional launch floor). Approval policy resolved at launch. OpenRig's default.
- review.reviewer: Can run shell commands as the launching user, subject to runtime and host policy. workspace-write (conditional launch floor). Approval policy resolved at launch. OpenRig's default.
- dogfood.tester: Can run shell commands as the launching user, subject to runtime and host policy. workspace-write (conditional launch floor). Approval policy resolved at launch. OpenRig's default.
- release.manager: Can run shell commands as the launching user, subject to runtime and host policy. acceptEdits (conditional launch floor). OpenRig's default.
- orch.lead: Can run shell commands as the launching user, subject to runtime and host policy. acceptEdits (conditional launch floor). OpenRig's default.
What they are told
- agents/conveyor-planner/guidance/role.md (plan.planner)
- agents/conveyor-planner/startup/context.md (plan.planner)
- agents/shared/skills/pm/requirements-writer (plan.planner)
- agents/shared/skills/pm/context-builder (plan.planner)
- agents/shared/skills/process/verification-before-completion (plan.planner, build.implementer, check.qa, review.reviewer, dogfood.tester, release.manager, orch.lead)
- CULTURE.md (plan.planner, build.implementer, check.qa, review.reviewer, dogfood.tester, release.manager, orch.lead)
- agents/implementer/guidance/role.md (build.implementer)
- agents/shared/skills/pods/development-team (build.implementer, check.qa)
- agents/shared/skills/process/systematic-debugging (build.implementer, check.qa, review.reviewer, dogfood.tester, orch.lead)
- agents/qa/guidance/role.md (check.qa)
- agents/shared/skills/process/agent-browser (check.qa)
- agents/shared/skills/process/dogfood (check.qa, dogfood.tester)
- agents/independent-reviewer/guidance/role.md (review.reviewer)
- agents/shared/skills/pods/review-team (review.reviewer)
- agents/factory-rsi-dogfood/guidance/role.md (dogfood.tester)
- agents/factory-rsi-release-manager/guidance/role.md (release.manager)
- agents/orchestrator/guidance/role.md (orch.lead)
- agents/shared/skills/pods/orchestration-team (orch.lead)
What else runs
- plugins: openrig-home:plugins/openrig-core, on if selected runtime supports and loads this resource (resolved at launch) (plan.planner, build.implementer, check.qa, review.reviewer, dogfood.tester, release.manager, orch.lead)
- runtime add-ons (settings, hooks, MCP servers): agents/shared/runtime/claude-settings.fragment.json, on if selected runtime supports and loads this resource (plan.planner, build.implementer, release.manager, orch.lead)
- runtime add-ons (settings, hooks, MCP servers): agents/shared/runtime/claude-mcp.fragment.json, on if selected runtime supports and loads this resource (plan.planner, build.implementer, release.manager, orch.lead)
- runtime add-ons (settings, hooks, MCP servers): agents/shared/runtime/claude-activity-hooks.json, on if selected runtime supports and loads this resource (plan.planner, build.implementer, release.manager, orch.lead)
- runtime add-ons (settings, hooks, MCP servers): agents/shared/runtime/codex-config.fragment.toml, on if selected runtime supports and loads this resource (check.qa, review.reviewer, dogfood.tester)
What it writes
Each depends on your configured destinations, what the runtime supports, and files already in place when it installs or launches.
- CLAUDE.md in the seat's working folder: add or update OpenRig's section (plan.planner, build.implementer, release.manager, orch.lead)
- .claude/skills/requirements-writer in the seat's working folder: project (plan.planner)
- .claude/skills/context-builder in the seat's working folder: project (plan.planner)
- .claude/skills/verification-before-completion in the seat's working folder: project (plan.planner, build.implementer, release.manager, orch.lead)
- .claude/plugins/shared:openrig-core in the seat's working folder: project (plan.planner, build.implementer, release.manager, orch.lead)
- .claude/settings.local.json in the seat's working folder: merge settings or managed hooks (plan.planner, build.implementer, release.manager, orch.lead)
- .mcp.json in the seat's working folder: merge MCP configuration fragment (plan.planner, build.implementer, release.manager, orch.lead)
- .claude/skills/development-team in the seat's working folder: project (build.implementer)
- .claude/skills/systematic-debugging in the seat's working folder: project (build.implementer, orch.lead)
- AGENTS.md in the seat's working folder: add or update OpenRig's section (check.qa, review.reviewer, dogfood.tester)
- .agents/skills/development-team in the seat's working folder: project (check.qa)
- .agents/skills/systematic-debugging in the seat's working folder: project (check.qa, review.reviewer, dogfood.tester)
- .agents/skills/verification-before-completion in the seat's working folder: project (check.qa, review.reviewer, dogfood.tester)
- .agents/skills/agent-browser in the seat's working folder: project (check.qa)
- .agents/skills/dogfood in the seat's working folder: project (check.qa, dogfood.tester)
- .codex/plugins/shared:openrig-core in the seat's working folder: project (check.qa, review.reviewer, dogfood.tester)
- config.toml in codex_home: merge managed configuration fragment (check.qa, review.reviewer, dogfood.tester)
- .agents/skills/review-team in the seat's working folder: project (review.reviewer)
- .claude/skills/orchestration-team in the seat's working folder: project (orch.lead)
- . in the install folder: unpack the team's files
Outside addresses
- 127.0.0.1
- api.example.com
- app.example.com
- arxiv.org
- claude.com
- code.claude.com
- cognition.com
- developers.openai.com
- example.com
- github.com
- localhost
- manus.im
- mcp.context7.com
- mcp.exa.ai
- openai.com
- platform.claude.com
- prod.example.com
- raw.githubusercontent.com
- remote-dev.example
- site-a.com
- site-b.com
- staging.example.com
- www.anthropic.com
- www.apache.org
- www.dbreunig.com
- www.google-analytics.com
- www.langchain.com
- www.openrig.dev
- www.philschmid.de
- www.trychroma.com
What you need
- claude-code (runtime: plan.planner, build.implementer, release.manager, orch.lead)
- claude-code: the user's own runtime account or provider configuration (login: plan.planner, build.implementer, release.manager, orch.lead)
- codex (runtime: check.qa, review.reviewer, dogfood.tester)
- codex: the user's own runtime account or provider configuration (login: check.qa, review.reviewer, dogfood.tester)
- OpenRig daemon >=0.6.6 (version)
- OpenRig CLI >=0.6.6 (version)
Not known before launch
- plan.planner: openrig-home:plugins/openrig-core: Resolved on the installing host at launch; contents were not read.
- build.implementer: openrig-home:plugins/openrig-core: Resolved on the installing host at launch; contents were not read.
- check.qa: openrig-home:plugins/openrig-core: Resolved on the installing host at launch; contents were not read.
- check.qa: agents/shared/runtime/codex-config.fragment.toml: Hook declarations could not be parsed; no code was executed.
- review.reviewer: openrig-home:plugins/openrig-core: Resolved on the installing host at launch; contents were not read.
- review.reviewer: agents/shared/runtime/codex-config.fragment.toml: Hook declarations could not be parsed; no code was executed.
- dogfood.tester: openrig-home:plugins/openrig-core: Resolved on the installing host at launch; contents were not read.
- dogfood.tester: agents/shared/runtime/codex-config.fragment.toml: Hook declarations could not be parsed; no code was executed.
- release.manager: openrig-home:plugins/openrig-core: Resolved on the installing host at launch; contents were not read.
- orch.lead: openrig-home:plugins/openrig-core: Resolved on the installing host at launch; contents were not read.
- host and native settings: Effective permissions, login, model availability, runtime versions, host catalogs/plugins and configured paths are not checked before launch.
- runtime behavior: Scripts and agents can compute other actions and addresses. Literal URL occurrences are not a prediction of network traffic.
- built-in and host integration: The installed OpenRig/runtime adds culture, onboarding, session identity and host context, and can provision runtime hook/settings files. Those bytes and effects are outside this archive.
Install
One command fetches exactly this configuration from GitHub and starts it.
Before you run it, you need:
- claude-code (runtime: plan.planner, build.implementer, release.manager, orch.lead)
- claude-code: the user's own runtime account or provider configuration (login: plan.planner, build.implementer, release.manager, orch.lead)
- codex (runtime: check.qa, review.reviewer, dogfood.tester)
- codex: the user's own runtime account or provider configuration (login: check.qa, review.reviewer, dogfood.tester)
- OpenRig daemon >=0.6.6 (version)
- OpenRig CLI >=0.6.6 (version)
rig up https://github.com/mvschwarz/openrig-world/tree/f1d1d8fa4beb3f4f4a2fd13b1cb1b3cbda21b560/rigs/launch/factory-rsi --cwd . --target ~/rigs/factory-rsiIt installs these three. The install prints the same three, so you can check they match:
- Commit
f1d1d8fa4beb- Configuration
build.implementer=claude-code,check.qa=codex,dogfood.tester=codex,orch.lead=claude-code,plan.planner=claude-code,release.manager=claude-code,review.reviewer=codex- Package digest
206dfc1795d2Covers the bundle's files, not your host's settings.
It unpacks the team's files into ~/rigs/factory-rsi, outside your project. Your project gets only the seats' own files.
npx skills add https://github.com/mvschwarz/openrig/tree/v0.7.20/skills/rigs --skill rigs --agent claude-codeRun it in your project folder, then type /rigs in Claude Code. Give it this team's page and tell it what you want to work on.
npx skills add https://github.com/mvschwarz/openrig/tree/v0.7.20/skills/rigs --skill rigs --agent codexRun it in your project folder, then type $rigs in Codex. Give it this team's page and tell it what you want to work on.
Needs Node.js 22 or 24 (on Apple silicon, 22), which provides npx. If your shell says npx isn't found, Node isn't installed or isn't on your PATH. How to install Node.js →
After you run it
Tell us how it went. Community reports appear on this page, apart from OpenRig's own testing.