菜单
10-node inference mesh router. Routes tasks to the right model based on poly_c tier, domain, and daily token budgets. Fallback chain automatic.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/EvezArt/evezart-openclaw --skill evez-mesh-router命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Syncs FIRE events and spine state to GitHub lord-evez666 (betweenness=0.92). Hash-chained commits. All FIRE events become repository_dispatch events.
Eigenvalue closure tracker. Receives Stripe webhooks, increments omega edges, tracks -0.358 → 0.0 eigenvalue closure. Hash-chained append-only spine.
EVEZ-OS cognitive layer for OpenClaw. Installs the full EVEZ Governance Lattice v0.2 into your OpenClaw workspace: CPF formula engine, FIRE event system, AEGIS threat monitor, DGM loop, revenue bridge, and 6-agent swarm.
基于 SOC 职业分类
正在显示 SKILL.md
| name | evez-mesh-router |
| version | 0.2.0 |
| description | 10-node inference mesh router. Routes tasks to the right model based on poly_c tier, domain, and daily token budgets. Fallback chain automatic. |
| tags | ["evez","mesh","inference","routing","llm"] |
Routes every LLM call to the right node based on task type and poly_c score. Tracks daily token budgets. Auto-fallback when limits hit.
| ID | Provider | Model | Daily Limit | Specialty |
|---|---|---|---|---|
| 0 | groq | llama-3.3-70b | 1K req | Speed — 300+ t/s |
| 1 | groq | llama-3.1-8b | 14.4K req | High-frequency |
| 2 | cerebras | llama-3.3-70b | 1K req | Fastest 70B on planet |
| 3 | gemini | gemini-2.0-flash | 1,500 req | Long context, multimodal |
| 4 | gemini | gemini-2.5-pro | 50 req | Deep analysis (save for poly_c≥5) |
| 5 | github | gpt-4o | 50 req | FIRE synthesis (save for poly_c≥5) |
| 6 | github | gpt-4.1 | 50 req | Code generation |
| 7 | github | o3 | 50 req | Reasoning (save for poly_c≥8) |
| 8 | sambanova | llama-405b | $5 credit | Max reasoning (MPPA-tier only) |
| 9 | openrouter | auto | 50 req/day | Free fallback |
def route(task_type, poly_c):
if poly_c >= 8.0: return ["github/o3", "sambanova/405b"]
if poly_c >= 5.0: return ["github/gpt-4o", "gemini-2.5-pro"]
if poly_c >= 0.9: return ["groq/70b", "cerebras", "gemini-flash"]
if task_type == "code": return ["github/gpt-4.1", "groq", "openrouter"]
return ["groq", "cerebras", "gemini-flash", "openrouter"]
/mesh status
→ groq/70b: 234/1000 req today
→ cerebras: 89/1000 req today
→ gemini-flash: 312/1500 req today
→ github/gpt-4o: 3/50 req today ← CONSERVE
→ github/o3: 0/50 req today ← CONSERVE (poly_c≥8.0 only)