Full orchestrator prompt: route Claude Code tasks across Haiku, Sonnet and the flagship
TawohAwa · x · 2026-11-11
The full paste-ready prompt from TawohAwa's thread, turning Claude into a project's lead orchestrator:
- Review the entire spec and break it into small, independently executable tasks with dependencies, acceptance criteria, and expected outcomes.
- Before each task, evaluate complexity, reasoning needs, risk, and scope to pick the right model instead of defaulting to the flagship.
- Haiku handles simple, well-defined changes; Sonnet handles standard implementation, refactoring, tests, and moderate debugging; the flagship is reserved for architecture decisions, hard debugging, ambiguous requirements, and system-wide changes.
- Independent tasks run in parallel where model-specific delegation is supported.
Paired with having the model write a detailed spec first, this forms a complete multi-model token-saving orchestration flow.
Related event: Claude Code Multi-Model Orchestration Workflow Saves Tokens(3 posts)→
More from coding & agent
- Stripe launches claimable sandboxes: frictionless payment integrations for AI agents — jeff_weinstein · 2026-11-12
- Databricks launches autonomous AI coworker Genie: 91% on enterprise tasks with half the tokens — Yuchenj_UW · 2026-11-12
- Loom founder launches Roma 2.0, a shared to-do list that orchestrates AI agents — thetripathi58 · 2026-11-12
- Developer proposes a 'blast radius' test for AI agents before building — Few-Garlic2725 · 2026-11-12
- Sosumi Eval: Apple Docs Access Lifts Coding Agent Pass Rates from ~33% to 94-98% — JordanMorgan10 · 2026-11-12
- CypherGuard fixes structurally wrong LLM Cypher queries, hitting 88.5% on CypherBench — JeremyCMorgan · 2026-11-12