Not in the top 1,000 by GitHub stars

OPCchuhai/OPC-koubo

橙哥出海开源的 Renoise Agent Skill:用 AI 做AI数字人口播短视频

History

Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.

View weekly changes

skills/digital-human-koubo/SKILL.md

Week of Nov 2, 2026 · Line counts unavailable

Before Not checked yet → After Network or broad shell

First observation → 57ad9920717a

No earlier version was observed for comparison.

Compare on GitHub

View on GitHubAdd to my skills

Verified commit 57ad9920717a3c286f96821bdf96348bff442bb8

13GitHub stars
Not availablestars in 7 days
1skill in this repository
12commits in 30 days

Heat

Insufficient history

Not enough history yet to measure heat. It appears after a few days of data.

Heat never changes the order by stars, and sponsorship never touches it. How heat works

Stars over time

Nov 2, 2026Nov 4, 202613

30-day change not yet available: not enough history.

Individual skills in this repository

How it works

digital-human-koubo

When your agent uses it

用 Renoise 做AI数字人口播 / talking-head 视频。用户提供自己的文案和自己有权使用的照片, 智能体连接官方 MCP(https://renoise.tv/api/public/mcp)后按大纲写稿、报价、生成口播。 尚未连接时,先判断用户用的是哪个 AI 产品,用本技能里对应的连接说明带他安装并登录。 尚未登录 Renoise 或需要注册时,给出本技能中的注册链接。不要改用别的注册地址,也不要改 MCP 地址。 触发:AI数字人口播、数字人讲这段、talking-head、口播视频、用数字人 Skill、橙哥出海数字人、连上 Renoise。 不用于:冒充真人、使用他人照片、填入上游或其他账号的 material ID。

What it loads into context

Frontmatter

194 tokens

Always available to the agent

SKILL.md

4,598 tokens

When the skill is activated

Reference documents

0 tokens

Loaded on request

Token counts are estimates. Scripts run separately; they are not counted as reference context.

The playbook

  1. 何时使用
  2. 注册链接(还没登录或需要注册时)
  3. 按产品连接 Renoise MCP
  4. 形象
  5. 流程大纲
  6. 红线

Scripts executed separately

  • scripts/renoise-manual-connect.sh

Files

  • SKILL.mdcore17,617 bytes
  • SKILL.mdcore17,617 bytes
  • SKILL.mdcore17,617 bytes
  • SKILL.mdcore17,617 bytes
  • SKILL.mdcore17,617 bytes
  • SKILL.mdcore17,617 bytes
  • persona.jsonasset1,039 bytes
  • scripts/renoise-manual-connect.shruns6,386 bytes
Read SKILL.md

# AI数字人口播

橙哥出海(Twitter/X:[OPCchuhai](https://x.com/OPCchuhai))的 Agent Skill 骨架。用固定数字人形象做口播 / talking-head 视频,生成走 Renoise。形象字段在同目录 `persona.json`,当前是占位,必须先用用户自己的照片重建。

本文件规定何时使用、注册链接、按产品连接 MCP,以及口播流程大纲。模型名、分辨率、积分单价、任务接口等生产细节以 **当前 Renoise 文档** 为准,再配合本 Skill 执行。不要把别的项目里的素材 ID、任务 ID 或人物照片抄进这里。

灵感来自 Berryxia / andyhuo520 的 [digital-human-koubo](https://github.com/andyhuo520/digital-human-koubo)(MIT)。这是橙哥出海(Twitter/X:[OPCchuhai](https://x.com/OPCchuhai))自行重写的技能包装,不是上游仓库的逐字节复制。技能本身免费,许可为 MIT。

## 何时使用

用户要做这类片子时启用本 Skill:

- AI数字人口播、数字人讲一段文案、talking-head / 口播短视频
- 带货、解读、教程、种草等「同一个人出镜说话」的成片
View the complete source on GitHub

Skills included

  • digital-human-koubo

Repository facts

License
MIT
Created
Nov 1, 2026
Last pushed
Nov 1, 2026
Forks
8
Open issues
0
Project activity
Active

Before you install

Network or broad shell

Automated signals, not a guarantee

Checked Nov 7, 2026

Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.

Show 5 notes
  • skills/digital-human-koubo/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • skills/digital-human-koubo/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • skills/digital-human-koubo/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • skills/digital-human-koubo/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
  • skills/digital-human-koubo/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
Show all 54 signals by file

skills/digital-human-koubo/SKILL.md

  • Network accessline 3
    description: >   用 Renoise 做AI数字人口播 / talking-head 视频。用户提供自己的文案和自己有权使用的照片,   智能体连接官方 MCP(https://renoise.tv/api/public/mcp)后按大纲写稿、报价、生成口播。
  • Network accessline 4
    用 Renoise 做AI数字人口播 / talking-head 视频。用户提供自己的文案和自己有权使用的照片,   智能体连接官方 MCP(https://renoise.tv/api/public/mcp)后按大纲写稿、报价、生成口播。   尚未连接时,先判断用户用的是哪个 AI 产品,用本技能里对应的连接说明带他安装并登录。
  • Network accessline 5
    智能体连接官方 MCP(https://renoise.tv/api/public/mcp)后按大纲写稿、报价、生成口播。
  • Network accessline 14
    橙哥出海(Twitter/X:[OPCchuhai](https://x.com/OPCchuhai))的 Agent Skill 骨架。用固定数字人形象做口播 / talking-head 视频,生成走 Renoise。形象字段在同目录 `persona.json`,当前是占位,必须先用用户自己的照片重建。
  • Network accessline 18
    灵感来自 Berryxia / andyhuo520 的 [digital-human-koubo](https://github.com/andyhuo520/digital-human-koubo)(MIT)。这是橙哥出海(Twitter/X:[OPCchuhai](https://x.com/OPCchuhai))自行重写的技能包装,不是上游仓库的逐字节复制。技能本身免费,许可为 MIT。
  • Network accessline 39
    https://www.renoise.ai/?utm_campaign=affiliate_program&utm_medium=affiliate&utm_source=rewardful&via=quantbang
  • Network accessline 41
    已经有账号、只是要接 MCP 时,不要把注册链接当成 MCP 地址。MCP 只用下一节里的 `https://renoise.tv/api/public/mcp`。
  • Network accessline 47
    https://renoise.tv/api/public/mcp
  • Network accessline 72
    "renoise": {       "type": "http",       "url": "https://renoise.tv/api/public/mcp"
  • Network accessline 73
    "type": "http",       "url": "https://renoise.tv/api/public/mcp"     }
  • Network accessline 74
    "url": "https://renoise.tv/api/public/mcp"
  • Network accessline 106
    脚本做的事:动态注册 OAuth 客户端 → 生成 PKCE + state → 在 `127.0.0.1:8765` 起本地回调 → 打印 `https://renoise.tv/mcp-connect?...` 授权链接让用户打开 → 用授权码换 token(**必须带浏览器 User-Agent,否则 Cloudflare 返回 403/1010 拦截**)→ token 存 `~/.renoise-mcp/token.json`(600 权限)→ 自动调用 `get_
  • Network accessline 108
    之后所有 MCP 调用走脚本生成的通道,端点仍是官方的 `https://renoise.tv/api/public/mcp`:
  • Network accessline 128
    安装:打开 https://chatgpt.com/plugins,添加一个名为 Renoise 的自定义 MCP 服务器,地址填 https://renoise.tv/api/public/mcp,登录方式选 OAuth。如果没有添加的选项,要先在 Settings › Apps › Advanced settings 打开 Developer mode。
  • Network accessline 142
    安装:打开 https://claude.ai/customize/connectors?modal=add-custom-connector&connectorName=Renoise&connectorUrl=https%3A%2F%2Frenoise.tv%2Fapi%2Fpublic%2Fmcp。Claude 的连接器设置会自动填好 Renoise 和地址,点 Add 即可。在 claude.ai 添加的连接器,Claude 桌面版里也能用。
  • Network accessline 158
    安装:运行 `openclaw plugins install @renoise/plugin` 安装 Renoise 插件,跟我确认后再运行 `openclaw mcp set renoise '{"url":"https://renoise.tv/api/public/mcp","transport":"streamable-http","auth":"oauth"}'` 添加 Renoise 服务器(传输方式必须是 streamable-http)。完成后让我开一个新会
  • Network accessline 172
    安装:豆包工作的 HTTP 插件走不完 OAuth 登录,所以 Renoise 要通过开源的 mcp-remote 桥接。在「插件」里新建一个名为 Renoise 的自定义插件,类型选 STDIO,命令填 npx,参数填 `-y mcp-remote@0.14.3 https://renoise.tv/api/public/mcp`,然后新开一个任务。
  • Network accessline 173
    登录:添加之前,我先在终端运行一次 `npx -y mcp-remote@0.14.3 https://renoise.tv/api/public/mcp`(需要 Node.js),在打开的页面登录,看到授权成功后按 Ctrl+C。登录状态会保存在这台电脑上。
  • Network accessline 186
    安装:千问办公的 HTTP 连接器走不完 OAuth 登录,所以 Renoise 要通过开源的 mcp-remote 桥接。在「扩展 › 连接器」手动添加一个名为 Renoise 的连接器,类型选 STDIO,命令填 `npx -y mcp-remote@0.14.3 https://renoise.tv/api/public/mcp`,然后新开一个对话。
  • Network accessline 187
    登录:添加之前,我先在终端运行一次 `npx -y mcp-remote@0.14.3 https://renoise.tv/api/public/mcp`(需要 Node.js),在打开的页面登录,看到授权成功后按 Ctrl+C。登录状态会保存在这台电脑上。
  • Network accessline 201
    mcp_servers:   renoise:     url: "https://renoise.tv/api/public/mcp"
  • Network accessline 202
    renoise:     url: "https://renoise.tv/api/public/mcp"     auth: oauth
  • Network accessline 203
    url: "https://renoise.tv/api/public/mcp"
  • Network accessline 220
    安装:添加一个名为 Renoise 的远程 MCP 服务器,地址 https://renoise.tv/api/public/mcp(Streamable HTTP)。它用 OAuth 登录,不需要 API key,也不用填请求头。
  • Network accessline 253
    - 注册引导只使用本技能里的那条注册链接,只当作给还没登录的人用的注册链接。 - MCP 只用 `https://renoise.tv/api/public/mcp`。
  • Network accessline 254
    - MCP 只用 `https://renoise.tv/api/public/mcp`。

skills/digital-human-koubo/scripts/renoise-manual-connect.sh

  • Executable scriptline 1
    #!/usr/bin/env bash
  • Network accessline 5
    # token 按 0 秒有效期写入,导致授权成功但工具数恒为 0(expired-without-refresh)。 # 本脚本绕过该客户端判断:动态注册 OAuth 客户端 → PKCE → 本地回环回调 → 换 token → # 直接以 Bearer 调用官方端点 https://renoise.tv/api/public/mcp(地址固定,不要改)。
  • Network accessline 6
    # 本脚本绕过该客户端判断:动态注册 OAuth 客户端 → PKCE → 本地回环回调 → 换 token → # 直接以 Bearer 调用官方端点 https://renoise.tv/api/public/mcp(地址固定,不要改)。 #
  • Network accessline 7
    # 直接以 Bearer 调用官方端点 https://renoise.tv/api/public/mcp(地址固定,不要改)。
  • Network accessline 14
    MCP_URL="https://renoise.tv/api/public/mcp"
  • Network accessline 15
    OAUTH_META="$MCP_URL/.well-known/oauth-protected-resource"
  • Network accessline 16
    REGISTER_URL="https://renoise.tv/api/public/mcp-oauth/register"
  • Network accessline 17
    TOKEN_URL="https://renoise.tv/api/public/mcp-oauth/token"
  • Network accessline 18
    AUTHZ_URL="https://renoise.tv/mcp-connect"
  • Network accessline 34
    echo "[2/6] 注册本地 OAuth 客户端..." CLIENT_JSON=$(curl -sS --max-time 20 -X POST "$REGISTER_URL" \
  • Network accessline 35
    CLIENT_JSON=$(curl -sS --max-time 20 -X POST "$REGISTER_URL" \
  • Network accessline 36
    -H "Content-Type: application/json" -H "User-Agent: $UA" \   -d "{\"client_name\":\"renoise-manual\",\"redirect_uris\":[\"http://127.0.0.1:$PORT/callback\"],\"grant_types\":[\"authorization_code\"],\"response_types\":[\"code\"],\"token_endp
  • Network accessline 37
    -d "{\"client_name\":\"renoise-manual\",\"redirect_uris\":[\"http://127.0.0.1:$PORT/callback\"],\"grant_types\":[\"authorization_code\"],\"response_types\":[\"code\"],\"token_endpoint_auth_method\":\"none\"}")
  • Dangerous shell operationline 59
    rm -f "$WORKDIR/callback.json"
  • Network accessline 66
    AUTH_URL="$AUTHZ_URL?response_type=code&client_id=$CLIENT_ID&redirect_uri=http%3A%2F%2F127.0.0.1%3A$PORT%2Fcallback&scope=$SCOPE&state=$STATE&code_challenge=$CHALLENGE&code_challenge_method=S256"
  • Network accessline 67
    echo "[3/6] 请在浏览器打开下面这条链接,登录 Renoise 并授权:" echo echo "  $AUTH_URL"
  • Network accessline 68
    echo echo "  $AUTH_URL" echo
  • Network accessline 69
    echo "  $AUTH_URL"
  • Network accessline 84
    echo "[5/6] 换取 access token..." HTTP_CODE=$(curl -sS --max-time 30 -X POST "$TOKEN_URL" \
  • Network accessline 85
    HTTP_CODE=$(curl -sS --max-time 30 -X POST "$TOKEN_URL" \
  • Network accessline 87
    --data-urlencode "grant_type=authorization_code" \   --data-urlencode "code=$CODE" \   --data-urlencode "redirect_uri=http://127.0.0.1:$PORT/callback" \
  • Network accessline 88
    --data-urlencode "code=$CODE" \   --data-urlencode "redirect_uri=http://127.0.0.1:$PORT/callback" \   --data-urlencode "client_id=$CLIENT_ID" \
  • Network accessline 89
    --data-urlencode "redirect_uri=http://127.0.0.1:$PORT/callback" \
  • Network accessline 103
    [Credential-related source omitted]
  • Access to secretsline 103
    [Credential-related source omitted]
  • Network accessline 104
    [Credential-related source omitted]
  • Dangerous shell operationline 105
    [Credential-related source omitted]
  • Network accessline 105
    curl -sS --max-time "${MCP_TIMEOUT:-120}" -X POST "https://renoise.tv/api/public/mcp" \

How the signals fit together

Facts from a static read of the files at the pinned commit. They describe what the code does, not why.

Signal confidence: 54 facts

and 4 more

Works with

Agent compatibility, from the files in this repository
AgentStatusEvidenceInstall to
Claude CodeCompatibleStandard SKILL.md skills/digital-human-koubo/SKILL.md.claude/skills/<skill-name>/
CodexCompatibleStandard SKILL.md skills/digital-human-koubo/SKILL.md.agents/skills/<skill-name>/
CursorCompatibleStandard SKILL.md skills/digital-human-koubo/SKILL.md.cursor/skills/<skill-name>/
Gemini CLICompatibleStandard SKILL.md skills/digital-human-koubo/SKILL.md.gemini/skills/<skill-name>/
GitHub CopilotCompatibleStandard SKILL.md skills/digital-human-koubo/SKILL.md.github/skills/<skill-name>/
OpenCodeCompatibleStandard SKILL.md skills/digital-human-koubo/SKILL.md.opencode/skills/<skill-name>/
  • Native The repository ships a manifest, folder or frontmatter field for this agent.
  • Compatible A standard SKILL.md that this agent can read.
  • Unknown No evidence either way.

Derived automatically from paths, manifests and frontmatter. Not tested on each agent.