● 社区
nyldn
🔑 需要你的密钥
Claude Octopus
作者 nyldn · nyldn/claude-octopus
Run up to 8 models in parallel on the same task; surface where they disagree — that's where your blindspot is.
Octopus is a multi-model orchestration skill: define a task, it dispatches to N model providers, then highlights consensus and disagreement. Useful for research, design review, and plan critique where one model can have a confident-but-wrong opinion.
为什么要用
核心特性
- Up to 8 models in parallel
- Disagreement highlighting
- Cost guard (caps spend per run)
- Supports Anthropic, OpenAI, Google, xAI, etc.
- Output is a single decision-aid Markdown
使用场景
实战用法： Claude Octopus
何时使用： You're about to commit to an architecture and want a second (and third, and fourth) opinion.
步骤
-
Brief
Provide the design doc + question.✓ 已复制
→ Octopus dispatches to N models
-
Read disagreements
Skim only where models disagree.✓ 已复制
→ Real risks surface
-
Decide
Resolve disagreements with a focused follow-up.✓ 已复制
→ Documented decision
结果： Decision made with explicit awareness of where models disagree.
组合
与其他 MCP 搭配，撬动十倍杠杆
Use Octopus on the proposed plan, then flow-next executes the agreed-on version
Combine claude-octopus with flow-next-skill: Use Octopus on the proposed plan, then flow-next executes the agreed-on version✓ 已复制
成本与限制
运行它的成本
- API 配额
- Per-provider limits
- 每次调用 Token 数
- N × normal
- 费用
- Pay-per-call across providers
- 提示
- Set the cost cap; use cheap models for the panel and one premium
安全
权限、密钥、影响范围
最小权限： multi-provider keys
凭据存储： API keys per provider in env
数据出站： Each configured provider
切勿授予： ship customer data through every provider blindly — review egress
故障排查
常见错误与修复
One provider 401s
Octopus continues with the rest; log shows which key
替代方案
Claude Octopus 对比其他方案
| 替代方案 | 何时用它替代 | 权衡 |
|---|
| Single-model with self-critique | Cost-sensitive | Misses provider-specific blindspots |