agentsclimarketplace

Chanjing customised person

Skill chanjing-ai/chan-skills/skills/chanjing-customised-person

蝉系openclaw 技能库,聚焦电商内容创作,提供实用的 AI 工具及技能包

Install
npx -y skills add chanjing-ai/chan-skills --skill chanjing-customised-person

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 17 stars17 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use Chanjing customised person APIs to create, inspect, list, poll, and delete custom digital humans from uploaded source videos.

SKILL.md

7.3 KB, ~2.1k tokens by cl100k_base, as published. Nobody here has run it

Chanjing Customised Person

功能说明

从本机上传源视频创建蝉镜定制数字人,并支持列表、查询、轮询与删除;流程与 Open API 一致:获取上传 URLcreate_upload_url)→ 对返回的 sign_url 发起 PUT 上传正文 → 创建/列出/获取/轮询/删除。仅使用蝉镜 Open API 与接口返回的上传端点,引入其它云厂商凭证。脚本依赖 ffmpeg/ffprobe。环境变量均可选(默认 ~/.chanjing 与官方 API 基址)。凭据、令牌持久化、网络边界见 manifest.yaml

运行依赖

  • python3 与同仓库 scripts/*.py;可读本地视频文件路径并上传

环境变量与机器可读声明

  • 环境变量键名与说明:manifest.yamlenvironment 段)及本文
  • 变量键名、凭据模型、合规 permissionsclientPermissionsagentPolicymanifest.yaml

使用命令

  • ClawHub(slug 以注册表为准):clawhub run chanjing-customised-person
  • 本仓库python skills/chanjing-customised-person/scripts/create_person.py …(见 Standard Workflow

登记与审稿(单一事实来源)

英文登记、主凭据、primaryEnv 省略、可选 env、敏感字段与出站边界:以同目录 manifest.yaml 为准(与 credential_hintdescription frontmatter 一致)。本篇从 When to Use 起写业务能力与流程,勿与 manifest.yaml 重复维护长篇对表

When to Use This Skill

当用户要做这些事时使用本 Skill:

  • 上传真人源视频,创建蝉镜定制数字人
  • 查询定制数字人列表或单个形象详情
  • 轮询定制数字人制作进度
  • 删除不再需要的定制数字人

如果需求是“拿已有数字人去合成口播视频”,优先使用 chanjing-video-compose
如果需求是“上传真人视频做对口型驱动”,优先使用 chanjing-avatar

Preconditions

执行本 Skill 前,必须先通过 chanjing-credentials-guard 完成 AK/SK 与 Token 校验。

本 Skill 与 guard 共用:

  • ~/.chanjing/credentials.json
  • https://open-api.chanjing.cc

无凭证时,脚本会自动打开蝉镜登录页,并提示配置命令。

审阅与安全(凭据与边界)

Purpose / Credentials / Persistence / Network 相关的逐项说明见 manifest.yaml(含 permissions.network_mode: opensignUrlPutNote)。以下仅 SKILL 正文补充

  • credentials.json:脚本会读取本地凭据;刷新后的 access_token / expire_in 写回同一文件预期行为。须限制目录/文件权限(如目录 0700、文件 0600),并确保永不提交版本库(见 manifest doNotCommitToVcs)。
  • sign_url 与 PUT 主机upload_file.py / get_upload_url.py 先请求 open-api.chanjing.cccreate_upload_url,再向响应中的 sign_url 发起 HTTPS PUT;该 URL 的主机常为蝉镜侧对象存储或 CDN,可能不在固定主机白名单中,清单中已改为 network_mode: open 以如实声明。
  • create_person.py --callback:若传入 URL,远程服务可能向该端点 POST 任务结果,载荷可能含状态与资源引用等敏感信息;须自行信任该端点并承担出站与数据暴露风险。
  • 本地视频:仅按用户给出的路径读取视频字节并上传到 API 指定端点;不扫描其它无关系统路径。

Standard Workflow

  1. 上传本地源视频,获取 file_id(推荐 upload_file.py:内部为 GET …/create_upload_urlPUT sign_url(正文为文件)→ 轮询 file_detail 至就绪;亦可分步用 get_upload_url.py + 自行 PUT + 轮询)
  2. 调用 create_person.py 创建定制数字人任务,得到 person_id
  3. 调用 poll_person.py 轮询直到成功,得到 preview_url,或用 get_person --field audio_man_id 拿到声音 id
  4. 如需批量查看历史形象,用 list_persons.py
  5. 如需清理资源,用 delete_person.py

Covered APIs

本 Skill 当前覆盖:

  • GET /open/v1/common/create_upload_url
  • GET /open/v1/common/file_detail
  • POST /open/v1/create_customised_person
  • POST /open/v1/list_customised_person
  • GET /open/v1/customised_person
  • POST /open/v1/delete_customised_person

Scripts

脚本目录:

  • skills/chanjing-customised-person/scripts/
脚本说明
_auth.py读取凭证、获取或刷新 access_token
get_upload_url.py调用 create_upload_url,输出 sign_urlmime_typefile_id 等 JSON(需自行 PUT)
upload_file.pycreate_upload_url + HTTPS PUT sign_url + 轮询 file_detail,输出 file_id
create_person.py创建定制数字人任务,输出 person_id;可选 --callback(服务端可能向该 URL POST 敏感任务载荷)
list_persons.py列出定制数字人形象
get_person.py获取单个数字人详情,默认输出 JSON
poll_person.py轮询形象详情直到完成,默认输出 preview_url
delete_person.py删除定制数字人,输出被删除的 person_id

Usage Examples

示例 1:从本地视频创建定制数字人

FILE_ID=$(python3 skills/chanjing-customised-person/scripts/upload_file.py \
  --file ./source.mp4)

PERSON_ID=$(python3 skills/chanjing-customised-person/scripts/create_person.py \
  --name "演示数字人" \
  --file-id "$FILE_ID" \
  --train-type figure)

python3 skills/chanjing-customised-person/scripts/poll_person.py --id "$PERSON_ID"

示例 2:查看完整详情

python3 skills/chanjing-customised-person/scripts/get_person.py \
  --id "C-ef91f3a6db3144ffb5d6c581ff13c7ec"

示例 3:列出与删除

python3 skills/chanjing-customised-person/scripts/list_persons.py

python3 skills/chanjing-customised-person/scripts/delete_person.py \
  --id "C-ef91f3a6db3144ffb5d6c581ff13c7ec"

Output Convention

默认不自动下载任何预览视频或封面图:

  • create_person.py 输出 person_id
  • poll_person.py 输出 preview_url,便于继续预览或保存
  • 只有在用户明确要求时,才应把返回的资源 URL 另存到本地

如果后续需要落盘预览资源,建议使用:

  • outputs/customised-person/

Additional Resources

更多接口细节与触发样例见:

  • skills/chanjing-customised-person/reference.md
  • skills/chanjing-customised-person/examples.md

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.