docs: say the new team names in shipped text

[mvschwarz]

The built-in teams are now starter, workshop (a rig bundle) and factory, plus the specialist teams code-review, research and pm (#864). This updates the shipped text that still named the old teams, so every page, skill, help string and onboarding asset offers the same teams the kernel operator does (#862).

What changed

• README.md and docs/reference/getting-started.md: the team table is starter, workshop and factory. Commands use rig up starter, dev-build@starter and dev-review@starter. The permissive-operation example uses a copied starter spec, and its Codex profile goes on dev.review, starter's Codex seat.
• docs/reference/rig-spec.md: the shipped-examples table lists starter, factory, code-review, research, pm and secrets-manager. A stale "implementation-pair skill" line in an example spec now says "Read your guidance and begin."
• docs/reference/chain-file-convention.md (factory ships the topology defaults now) and docs/reference/sdlc-conventions.md (pm).
• packages/cli/src/commands/setup.ts: the next steps name starter, workshop and factory, and say the operator adapts a team for one provider. packages/cli/test/setup.test.ts is updated to match.
• packages/cli/src/commands/context.ts help examples and scripts/vm-bootstrap/two-daemon-start.sh use factory.
• Skills: openrig-user (a 0.6.6 note beside its v0.3.x history), openrig-architect, and openrig-software-factory with its worked example (addresses, rig name, and the builder's --runtime claude-code).
• context-packs-src/onboarding-width/public-what-you-can-do.md: the first-change paragraph.
• context7.json: install prerequisites.

The one fact that isn't a rename: first-project was two Codex agents, and starter is a Claude builder plus a Codex reviewer. So the text no longer tells a Codex-only reader to run first-project. It points them to the kernel operator, which writes an adapted copy of the team for their providers. first-project appears only as starter's old name.

Left on purpose

• conveyor as the workflow concept (workflow help, conveyor.yaml, "no implicit fixed-role review conveyor").
• The v0.3.x history line in openrig-user (with the 0.6.6 note beside it).
• Example names such as my-product-team, test fixtures, docs/as-built, release notes and the web UI.

Generated files: the skills/_canonical mirrors and the skill-edge digests were regenerated on this branch (20776fd), after a rebase onto #864. secrets-manager stays listed as a built-in team (e04515c), since it stays built in for 0.6.6.

Summary by CodeRabbit

• Documentation
  • Updated setup guides and examples to use the starter, workshop, and factory teams, with guidance for adapting to available AI tools.
  • Clarified built-in and specialist team options, launch steps, and seat names.
  • Updated examples to reflect the current team names and workflows.

[coderabbitai]

Note

Currently processing new changes in this PR. This may take a few minutes, please wait...

⚙️ Run configuration

• Configuration used: defaults
• Review profile: CHILL
• Plan: Advanced
• Run ID: 5d29b849-0ff4-4621-9691-24cee7b5162a

📥 Commits

Reviewing files that changed from the base of the PR and between d1e9d5e and e04515c.

⛔ Files ignored due to path filters (1)

• scripts/skill-edge-digests.generated.json is excluded by !**/*.generated.*

📒 Files selected for processing (18)

• README.md
• context7.json
• docs/reference/chain-file-convention.md
• docs/reference/getting-started.md
• docs/reference/rig-spec.md
• docs/reference/sdlc-conventions.md
• packages/cli/src/commands/context.ts
• packages/cli/src/commands/setup.ts
• packages/cli/test/setup.test.ts
• packages/daemon/assets/plugins/openrig-core/skills/openrig-user/SKILL.md
• packages/daemon/context-packs-src/onboarding-width/public-what-you-can-do.md
• packages/daemon/specs/agents/shared/skills/core/openrig-architect/SKILL.md
• packages/daemon/specs/agents/shared/skills/core/openrig-software-factory/SKILL.md
• packages/daemon/specs/agents/shared/skills/core/openrig-software-factory/references/worked-example.md
• scripts/vm-bootstrap/two-daemon-start.sh
• skills/_canonical/core/openrig-architect/SKILL.md
• skills/_canonical/core/openrig-software-factory/SKILL.md
• skills/_canonical/core/openrig-software-factory/references/worked-example.md

 ____________________________________________
< Forcing bugs to surrender unconditionally. >
 --------------------------------------------
  \
   \   (\__/)
       (•ㅅ•)
       / 　 づ

✨ Finishing Touches

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[openrig-review]

Approved at e04515c. Every shipped page, skill, help string and onboarding asset that named the old built-in teams now offers starter (a Claude builder and a Codex reviewer), workshop (a rig bundle from /rigs) and factory, plus the code-review, research and pm shelf, with the approved seat addresses. first-project appears once per file as starter's old name, and a Codex-only reader is pointed to the kernel operator, which adapts the team, instead of being told to swap names. secrets-manager stays listed as built in. The skill mirrors and edge digests are regenerated for the changed shared skills, and setup's printed next steps and their test are updated together.

— dev60-planner@v-openrig-build