add-minimax-provider
为 Hermes 配置 MiniMax 作为模型源。MiniMax 提供两种接入方式:API Key 直连(openai-completions 协议)和 OAuth 门户(anthropic-messages 协议)。包含 provider 注册、模型定义、别名配置、fallback…
Skill by jleechanorg
Chalk Labs
Need help deploying add-minimax-provider?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/jleechanorg-add-minimax-provider
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: backend, security
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/add-minimax-provider
What's inside
- Skill name
add-minimax-provider- Description
- 为 Hermes 配置 MiniMax 作为模型源。MiniMax 提供两种接入方式:API Key 直连(openai-completions 协议)和 OAuth 门户(anthropic-messages 协议)。包含 provider 注册、模型定义、别名配置、fallback 链接入和验证的完整流程。当管理员说想"加 MiniMax"、"配 minimax"、"接入 MiniMax 模型"、"加海螺模型"、"配 M2.1"时使用此 skill。
- License
- MIT
- Instructions
- 417 lines in SKILL.md · view on GitHub
Before installing
- Instructions call the network
Files (4)
- origin.json · .clawhub/
- README.md
- SKILL.md
- _meta.json
SKILL.md (first lines)
# 配置 MiniMax Provider(MiniMax 大模型接入)
MiniMax 是国产大模型服务商,提供 MiniMax-M2.1 系列模型。Hermes 支持两种接入方式:
| Provider | 接入方式 | API 协议 | Base URL | 适用场景 |
|----------|----------|----------|----------|----------|
| `minimax` | API Key 直连 | `openai-completions` | `api.minimaxi.com/v1` | 标准付费用户 |
| `minimax-portal` | OAuth 门户 | `anthropic-messages` | `api.minimaxi.com/anthropic` | 免费/门户用户 |
**为什么有两个 provider?**
- **API Key 直连** (`minimax`):使用标准 OpenAI 兼容协议,需要付费 API Key,按量计费
- **OAuth 门户** (`minimax-portal`):使用 Anthropic Messages 协议 + OAuth 认证,适合有门户账号的用户(可能有免费额度)
两种方式访问的是同一个模型,但协议和认证方式不同,所以必须分开配置。
如果觉得这个 Skill 有用,欢迎通过邀请链接注册 MiniMax Coding Plan(9 折优惠 + Builder 权益):
https://platform.minimaxi.com/subscribe/coding-plan?code=2vNMQFJrZt&source=link
---
## 零起步冷启动
**还没有任何模型可用?** 可以用 Qwen Coder(免费)作为冷启动模型,然后让它帮你配置 MiniMax。
### 步骤
1. **先配好 Qwen Coder**(免费,无需 API Key,OAuth 登录即可):
- 在 Hermes 中添加 `qwen-portal` provider
- 用 Qwen Coder 作为 primary 模型启动系统
2. **让 Qwen Coder 帮你配 MiniMax**:
- 在聊天中告诉 Agent:"帮我配置 MiniMax 模型"
- Agent 会自动加载本 skill 并按步骤执行
- 配置完成后,将 primary 切换为 MiniMax
3. **切换主力模型**:
```
/model Minimax
```
> 这就是 Hermes 的"自举"能力——用一个免费模型启动系统,再用它来配置更强的模型。
---
## 可用模型
| 模型 ID | 名称 | Context | reasoning | 说明 |
|---------|------|---------|-----------|------|
| `MiniMax-M2.1` | MiniMax M2.1 | 200K | ❌ | 主力模型,综合能力强 |
| `MiniMax-M2.1-lightning` | MiniMax M2.1 Lightning | 200K | ❌ | 轻量快速版(仅 portal 确认可用) |
> **注意**:MiniMax 可能随时上线新模型。添加前务必先测试可用性。
---
## 前置条件
### 方式一:API Key 直连(推荐)
| 项目 | 说明 |
|------|------|
| 国内MiniMax 账号 | 在 [platform.minimaxi.com](https://platform.minimaxi.com) 注册 |
| API Key | 格式为 `sk-cp-...` 的密钥,在控制台 → API Keys 页面创建 |
| 余额 | 确保账户有足够余额 |
### 方式二:OAuth 门户
| 项目 | 说明 |
|------|------|
| MiniMax 账号 | 同上 |
| OAuth 配置 | 在 Hermes 中配置 `apiKey: "minimax-oauth"`,Hermes 会自动处理 OAuth 流程 |
---
## 第一步:测试模型可用性
**这一步不能跳过。** 先确认模型能调通再加配置。
### 1A. 测试 API Key 直连(openai-completions 协议)
```bash
curl -s --max-time 15 https://api.minimaxi.com/v1/chat/completions \
-H "Authorization: Bearer <你的API_KEY>" \
-H "Content-Type: application/json" \
-d '{"model":"MiniMax-M2.1","messages":[{"role":"user","content":"hi"}],"max_tokens":10}'
```
如果返回正常的 JSON 响应(含 `choices`)= 可用。
### 1B. 测试 OAuth 门户(anthropic-messages 协议)
OAuth 门户无法直接用 curl 测试(需要 OAuth token 流程)。配置好后通过 Hermes 实际发消息验证。
### 1C. 测试 Lightning 模型
```bash
curl -s --max-time 15 https://api.minimaxi.com/v1/chat/completions \
-H "Authorization: Bearer <你的API_KEY>" \
-H "Content-Type: application/json" \
-d '{"model":"MiniMax-M2.1-lightning","messages":[{"role":"user","content":"hi"}],"max_tokens":10}'
```
> 如果返回错误,说明该模型可能仅在 portal 端可用,或尚未对你的账户开放。
---
## 第二步:添加 Provider
在 `~/.smartclaw/config.yaml` 的 `models.providers` 下添加 provider。根据你的接入方式选择一种或两种都配。
### 2A. 添加 minimax(API Key 直连)
```json
"minimax": {
"baseUrl": "https://api.minimaxi.com/v1",
"apiKey": "<你的API_KEY>",
"api": "openai-completions",
"authHeader": true,
"models": [From jleechanorg/jleechanbrain, MIT.
More from jleechanorg/jleechanbrain
- 0.559score
agent-agent-mistakes
jleechanorg
Build agent memory patterns that prevent the same mistakes — using session replay, error tracking, and a self-correcting knowledge base.
SkillBFreshness: B★ 2Claude Code +9 - 0.559score
ao-babysit
jleechanorg
Babysit an active AO worker tmux session — poll it, detect stalls/deaths, steer it back on track, post progress to Slack.
SkillBFreshness: B★ 2Claude Code +9 - 0.557score
advice
jleechanorg
Token-efficient second opinion slash command /advice.
SkillBFreshness: B★ 2Claude Code +9 - 0.557score
claw-dispatch
jleechanorg
Use when dispatching work through the Hermes gateway with /claw, especially when the task may resolve slash commands or hand off into AO worker orchestration.
SkillBFreshness: B★ 2Claude Code +9 - 0.554score
social-poster
jleechanorg
Draft, stage, and publish social-media posts for Reddit, Hacker News, Twitter/X, Mastodon, LinkedIn, Threads, Facebook, Instagram, and Dev.to.
SkillBFreshness: B★ 2Claude Code +9 - 0.553score
ao-dispatch-mechanics
jleechanorg
AO spawn flags, model-switch, partial-work recovery.
SkillBFreshness: B★ 2Claude Code +9 - 0.553score
half-stop-user-re-ping-signal
jleechanorg
Detect when user re-sends their own prior message verbatim.
SkillBFreshness: B★ 2Claude Code +9 - 0.553score
agent-autonomy-failure-classes
jleechanorg
Catalog of agent autonomy failures with SOUL.md fixes.
SkillBFreshness: B★ 2Claude Code +9 - 0.550score
hermes-harness
jleechanorg
jleechanbrain / Hermes-specific harness — gateway, canary, deploy, lane backlog, staging vs prod.
SkillBFreshness: B★ 2Claude Code +9 - 0.549score
agent-pipeline-render
jleechanorg
Render artifacts from agent-swarm repos without LLM.
SkillBFreshness: B★ 2Claude Code +9 - 0.544score
agento
jleechanorg
Delegate coding change tasks to Agent-Orchestrator (AO).
SkillBFreshness: B★ 2Claude Code +9 - 0.543score
hermes-single-dir-pipeline
jleechanorg
Operate Hermes after the single-directory runtime migration.
SkillBFreshness: B★ 2Claude Code +9
Signals
- Stars
- 2
- Forks
- 0
- Open issues
- 1
- Last commit (this folder)
- Jun 21, 2026 (4 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/jleechanorg/jleechanbrain
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from jleechanorg
- 0.559score
ao-babysit
jleechanorg
Babysit an active AO worker tmux session — poll it, detect stalls/deaths, steer it back on track, post progress to Slack.
SkillBFreshness: B★ 2Claude Code +9 - 0.559score
agent-agent-mistakes
jleechanorg
Build agent memory patterns that prevent the same mistakes — using session replay, error tracking, and a self-correcting knowledge base.
SkillBFreshness: B★ 2Claude Code +9 - 0.557score
claw-dispatch
jleechanorg
Use when dispatching work through the Hermes gateway with /claw, especially when the task may resolve slash commands or hand off into AO worker orchestration.
SkillBFreshness: B★ 2Claude Code +9 - 0.557score
advice
jleechanorg
Token-efficient second opinion slash command /advice.
SkillBFreshness: B★ 2Claude Code +9 - 0.554score
disk-root-cause
jleechanorg
Use when diagnosing why a disk is full, what consumes its capacity, what grew over time, or what may be safely reclaimed.
SkillAFreshness: AClaude Code +9 - 0.554score
social-poster
jleechanorg
Draft, stage, and publish social-media posts for Reddit, Hacker News, Twitter/X, Mastodon, LinkedIn, Threads, Facebook, Instagram, and Dev.to.
SkillBFreshness: B★ 2Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get add-minimax-provider running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
add-minimax-provider skill. Agent Mart. https://agentmart.dev/skills/jleechanorg-add-minimax-provider (last updated 2026-10-24)