Multi-agent strategy
The project goal is practical:
Make Plus feel like Pro.
Use the strongest model for judgment. Use cheaper or faster workers for scoped execution. Keep Codex in charge.
Operating model
| Layer | Role |
|---|---|
| Brain | Codex plans, routes, reviews, and decides |
| Hands | Claude Code workers perform scoped tasks |
| Router | CCSwitch provides local profiles and models |
| Ledger | Each run leaves metadata, prompts, stdout, and stderr |
Default roles
| Role | Use it for |
|---|---|
requirements | Scope, non-goals, acceptance criteria |
architecture | Repo map, likely files, implementation plan, risk |
development | Main code development tasks |
testing | Test plan, edge cases, validation commands |
review | Findings, file references, maintainability |
performance | Runtime, IO, latency, resource use |
compatibility | Windows, macOS, Linux, shell, version risk |
documentation | Tutorials, FAQ, examples, onboarding |
automation | CI, release workflows, package checks |
security | Secrets, permissions, command risk, supply chain |
implementation | Scoped edits when write access is allowed |
ops | Deployment, logs, rollback, runtime risk |
multimodal | Image or mixed input work |
Four-phase workflow
- Parallel analysis: requirements, architecture, security, testing, and other relevant roles inspect the task.
- Cross-review: agents compare risk, scope, and plan quality.
- Execution: implementation runs only after the plan is stable and write scope is clear.
- Controller summary: Codex reviews logs, diffs, tests, and final output.
Generate the plan:
bash
python "$CC_ORCHESTRATOR_HOME/cc_orchestrator.py" workflow-plan "Ship a safe refactor"Cost rule
Do not spend the best model on every subtask.
Use it for:
- final judgment
- architecture
- risky review
- hard tradeoffs
Use worker models for:
- repo mapping
- test design
- documentation draft
- compatibility checks
- focused implementation after approval
This is the main reason the project exists.