skill-publish-sync

Skill Sync Tool (ClawHub + Tencent SkillHub)

Batch-sync locally developed Skills to three public platforms:

ClawHub — International general-purpose Skills community, enforces MIT-0 license
Tencent SkillHub — Skills community optimized for Chinese users, no license restrictions
Lenovo Open Platform — For the Lenovo AI agent ecosystem, uses @lenovo-open/skill-cli (Node.js); no mandatory license

Supports reading .gitignore to smartly ignore sensitive files and temporary files. The three platforms each maintain an independent allowlist file (allowlist-clawhub.yaml / allowlist-skillhub.yaml / allowlist-lenovo.yaml); the same skill can have its fields maintained independently across multiple files (one-to-many, no forced non-duplicate distribution).

Dependencies

System dependencies

Dependency	Purpose	Installation method
`rsync`	`prepare-publish.sh` Copy filtered files to the temporary publish directory	Bundled with macOS; Linux: `sudo apt-get install rsync`
`git`	`prepare-publish.sh` Use `git ls-files` to precisely match tracked files	Bundled with macOS; Linux: `sudo apt-get install git`
Python 3	Run the Tencent SkillHub CLI (Python script)	Bundled with macOS; Linux: `sudo apt-get install python3`
Node.js（≥ 18）+ `npm`	Run Lenovo `lenovoskill` CLI（`npx @lenovo-open/skill-cli` runs on the Node.js runtime)	Bundled with macOS or `brew install node`；Linux: `sudo apt-get install -y nodejs npm`

CLI tools (install on demand per target platform)

CLI	Purpose	Installation command	Required scenario
`clawhub`	Publish to ClawHub	See ClawHub official documentation	Only when publishing to ClawHub
`skillhub`	Publish to Tencent SkillHub	Official CLI,`skillhub self-upgrade` upgrade (must be ≥ 2026.7.29)	Only when publishing to SkillHub
`lenovoskill`（`npx @lenovo-open/skill-cli`）	Upload to Lenovo Open Platform	`npx @lenovo-open/skill-cli <cmd>`(no install required, recommended) or `npm i -g @lenovo-open/skill-cli`	Only when uploading to Lenovo Platform

⚠️ Version requirement (important)

Tencent skillhub CLI must be ≥ 2026.7.29 to support publish/login. Earlier versions (such as 2026.3.18) only have search/install, and cannot publish. Run skillhub self-upgrade to upgrade to the latest version; if unsure, first use skillhub self-upgrade --check-only to check whether a new version is available.

First-time installation checklist (SkillHub publishing)

# 1. 安装 CLI（macOS/Linux）
curl -fsSL https://skillhub-1388575217.cos.ap-guangzhou.myqcloud.com/install/install.sh | bash

# 2. 升级到支持发布的版本（≥ 2026.7.29）
skillhub self-upgrade

# 3. 验证（应含 publish/login 命令）
skillhub --help

# 4. 登录（token 从 https://skillhub.cn 个人设置获取，skh_ 前缀）
skillhub login --key skh_xxx

# 5. 确认身份
skillhub auth whoami

ClawHub enforces the MIT-0 license(no attribution required, commercial use allowed). Tencent SkillHub has no license restrictions, so CC-BY-NC / MIT / Apache etc. can all be published.

ClawHub license key terms(from the official Skill Format documentation, verified as of 2026-08):

"All skills published on ClawHub are licensed under MIT-0." —— All publishing is equivalent to MIT-0.

"Anyone may use, modify, and redistribute published skills, including commercially." —— Anyone may use them commercially.

"Attribution is not required." —— No attribution required.

"ClawHub does not support per-skill license overrides." —— Per-skill license overrides are not supported(frontmatter writes license: CC-BY-NC platform-level does not take effect).

"Do not add conflicting license terms in SKILL.md." —— Do not add conflicting license terms in SKILL.md。

Therefore (v1.7.1 three independent list architecture）：

Each platform independently maintains its own allowlist file: one each for ClawHub / SkillHub / Lenovo

MIT-licensed general-purpose tool skill → by default publish only to ClawHub (allowlist-clawhub.yaml）

Legal skills under restrictive licenses such as CC-BY-NC → by default publish only to SkillHub (allowlist-skillhub.yaml, ClawHub MIT-0 conflicts are not accepted)

Skills that want to be distributed additionally to the Lenovo ecosystem beyond ClawHub / SkillHub → register in allowlist-lenovo.yaml (no license restrictions, can be registered in duplicate with SkillHub / ClawHub, i.e. the same skill appearing across multiple files is completely allowed)

If the ClawHub slug is taken and cannot be published → move from allowlist-clawhub.yaml to allowlist-skillhub.yaml or allowlist-lenovo.yaml

i.e.three independent allowlist files, each independently maintains the skill list and fields for its own platform; the same skill can appear in multiple files at the same time (one-to-many), with no need to change the configuration of old skills for a new platform. See DECISIONS D-2026-08-12-02。

For ClawHub licenses see ClawHub Skill Format official documentation

Platform comparison

Dimension	ClawHub	Tencent SkillHub	Lenovo Open Platform
CLI tool	`clawhub`	`skillhub`(official CLI,`skillhub self-upgrade` upgrade;must be ≥ 2026.7.29 to support publish/login)	`lenovoskill`（`npx @lenovo-open/skill-cli`, Node.js runtime)
Login command	`clawhub login`	`skillhub login --key skh_xxx`	`lenovoskill login`(OAuth browser authorization)
Publish command	`clawhub publish <path> --slug --name --version --changelog`	`skillhub publish <path> [--version] [--changelog] [--dry-run]`	`lenovoskill package && lenovoskill push`(two steps: first create the zip then upload)
skill identifier	`--slug` + `--name`(passed in via the command line)	`slug` + `displayName`(prepare-publish.sh injects a temporary copy from `allowlist-skillhub.yaml` ; the source SKILL.md does not contain it)	`slug` + `displayName`(prepare-publish.sh injects a temporary copy from `allowlist-lenovo.yaml` ; the source SKILL.md does not contain it) + in-project `.skill-config.json`
Credential location	clawhub CLI default	local skillhub configuration	`~/.lenovoskill/auth.json`(600, OAuth dual token: Login Token + Biz Token) + `~/.lenovoskill/config.json`(default API)
Visibility	None	— (no such concept)	— (no such concept)
namespace	None	Yes(bound to the account, not passed in the publish command, recorded for traceability)	None(OAuth dual token, no namespace concept)
License	Mandatory MIT-0	Unrestricted	Unrestricted
Default registry/API	Built-in	`https://api.skillhub.cn`	`https://open.lenovomm.com`
Official tutorial	—	https://skillhub.cn/tutorials#publish-via-cli	https://open.lenovomm.com

Prerequisites

Required SKILL.md frontmatter fields

---
name: skill-name
description: 技能描述
version: "1.0.0"  # 推荐但不强制
homepage: https://github.com/cat-xierluo/legal-skills  # 自动设置
---

slug / displayName are not written into SKILL.md. The fields required by SkillHub are read from config/allowlist-skillhub.yaml ; the fields required by Lenovo Open Platform are read from config/allowlist-lenovo.yaml .Each platform has an independent allowlist, and fields are maintained independently(The same skill can appear in multiple files at once, one-to-many). Before publishing, prepare-publish.sh（--platform skillhub or --platform lenovo) automatically injects a temporary copy of the frontmatter. The source SKILL.md retains only the standard skill fields, decoupling platform metadata from the skill itself. ClawHub does not rely on frontmatter (it uses --slug/--name command-line arguments).

CLI Installation

For the complete installation method and version requirements of the CLI tool, see the "Dependencies" section at the beginning of this document.

Quickly verify installation:

clawhub --version      # ClawHub
skillhub --help        # SkillHub（确认含 publish/login 命令即为发布版）

Usage

1. Log in to the platform (first-time use)

ClawHub：

clawhub login

Tencent SkillHub(uses an API Token, format skh_xxx, obtained from SkillHub personal settings):

# token 从 https://skillhub.cn 个人设置获取（skh_ 前缀）
skillhub login --key skh_xxx

# 如需临时使用其他 token 覆盖登录态（不写入本地）
skillhub login --key skh_yyy

After logging in, the token is stored locally, and subsequent publish commands do not need to pass the token again. You can also temporarily override it with skillhub publish when --token skh_xxx .

2. Verify login status

# ClawHub
clawhub whoami

# 腾讯 SkillHub
skillhub auth whoami

3. Sync skills

ClawHub

Sync a single skill：

clawhub sync skills/<skill-name>

Sync all skills：

clawhub sync --all

Note:--all is subject to skills/skill-publish-sync/config/allowlist-clawhub.yaml . Only skills listed in that file will be synced to ClawHub. The same applies to SkillHub / Lenovo (each has its own independent whitelist file).

Tencent SkillHub

SkillHub recommends publishing skills one by one (specify the directory via the publish command; the identifier comes from the SKILL.md frontmatter, see "Single Skill Sync Workflow" below).

# 发布目录（自动打包）
skillhub publish ./my-skill

# 指定版本和 changelog
skillhub publish ./my-skill --version 1.2.0 --changelog "新增xxx"

# 先预检（只校验 + 打包，不实际发请求）
skillhub publish ./my-skill --dry-run

SkillHub uses slug + displayName(before publishing, prepare-publish.sh injects it into a temporary copy from config/allowlist-skillhub.yaml ; the source SKILL.md does not contain it) to identify the skill. The namespace is bound to the account (no need to specify it on the command line when publishing). There is no concept of visibility.

Interactive selection sync: The user can specify the list of skills to sync and the target platform, and I will execute the sync commands one by one.

Single Skill Sync Workflow

Use this workflow when you need to sync specified skills (rather than all of them). The process is common to both platforms; only the commands differ in the "Execute Publish" step.

Pre-checks

Check login status

clawhub whoami           # ClawHub
skillhub auth whoami     # 腾讯 SkillHub

Check whitelist
- Read skills/skill-publish-sync/config/allowlist-${PLATFORM}.yaml(platform-specific whitelist)
- Confirm that the target skill appears in the allowlist file for the corresponding platform (each platform has an independent file, with no platforms field)
- Independent file paths for the three platforms:
  - ClawHub → allowlist-clawhub.yaml
  - SkillHub → allowlist-skillhub.yaml
  - Lenovo → allowlist-lenovo.yaml
Check license(required only for ClawHub)
- Read the license field in the target skill's SKILL.md frontmatter
- ClawHub: only skills with the MIT license can be synced
- SkillHub / Lenovo: no license restrictions, all can be synced

Version detection

Compare two version numbers (read the record according to the target platform):

Source	Location	Format
New version	`skills/<skill-name>/SKILL.md` frontmatter's `version`	`"1.2.0"`
Recorded version	`skills/skill-publish-sync/config/sync-records.yaml` in `records.<skill>.platforms.<platform>.version`	`"1.1.0"`

Version comparison logic(semantic versioning):

new_version > recorded_version → 需要同步
new_version == recorded_version → 跳过（无变化）
new_version < recorded_version → 警告（版本回退？）
recorded_version 为 null → 需要同步（首次发布）

Execute sync

Step 1: Prepare the publish directory

# ClawHub（默认平台，可省略 --platform）
bash skills/skill-publish-sync/scripts/prepare-publish.sh skills/<skill-name>

# 腾讯 SkillHub
bash skills/skill-publish-sync/scripts/prepare-publish.sh --platform skillhub skills/<skill-name>

The script will create a temporary directory:

ClawHub：/tmp/clawhub-publish-<skill-name>
SkillHub：/tmp/skillhub-publish-<skill-name>

Step 2: Execute publishing

ClawHub(Use clawhub publish, explicitly specify slug and name):

clawhub publish /tmp/clawhub-publish-<skill-name> \
  --slug <skill-name> \
  --name "<Display Name>" \
  --version "<新版本号>" \
  --changelog "<变更说明>"

⚠️ You must specify --slug and --name

Temporary directory names may contain a prefix (such as clawhub-publish-), causing an incorrect slug at publish time

--slug <skill-name> Ensure the correct skill identifier is used

--name "<Display Name>" Ensure the correct name is displayed on ClawHub

Tencent SkillHub(Use skillhub publish, specify version and changelog):

skillhub publish /tmp/skillhub-publish-<skill-name> \
  --version "<新版本号>" \
  --changelog "<变更说明>"

Tencent identifies skills by slug + displayName: These two fieldsare not written into the source SKILL.md, but are read before publishing by prepare-publish.sh --platform skillhub from config/allowlist-skillhub.yaml and injected into the temporary copy's frontmatter (slug defaults to name, displayName takes the configured display_name). The namespace is bound to the account (no command-line specification is needed at publish time), and there is no concept of visibility. The version number and changelog are passed explicitly via --version / --changelog ; without --version , it is decided by the CLI. Before publishing, you can add --dry-run to perform only a pre-check.

Lenovo Open Platform(Use lenovoskill, first build the zip and then upload in two steps):

# 1. 准备：先在临时目录生成 .skill-config.json（联想 CLI 项目级配置）
cd /tmp/lenovo-publish-<skill-name>
# 源 skill 若已带 .skill-config.json,prepare-publish.sh 已一并复制,跳过 init;
# 否则跑一次 init 生成默认配置再按需编辑:
lenovoskill init

# 2. 打 zip
lenovoskill package
# → 生成 <skill-name>.zip

# 3. 上传 zip
lenovoskill push

Lenovo likewise reads slug + displayName from the SKILL.md frontmatter: These two fields are injected by prepare-publish.sh --platform lenovo into the temporary copy (sharing the same Python logic with SkillHub). At the same time, the Lenovo CLI also reads the project's .skill-config.json— the two coexist, and thedisplayName / slug field should remain consistent between the SKILL.md frontmatter and .skill-config.json . .skill-config.json RecommendedGenerate/copy only in the temporary directory, and do not commit it into the source skill (it is Lenovo CLI's private adaptation layer and has nothing to do with the standard skill structure).

Step 3: Update the sync record

Update skills/skill-publish-sync/config/sync-records.yaml, and write the record under the corresponding platform:

records:
  <skill-name>:
    platforms:
      clawhub:                      # 或 skillhub / lenovo
        version: "<新版本号>"
        last_sync: "<ISO 8601 时间>"
        git_hash: "<当前 commit hash>"
        status: synced
        changelog_summary: "<变更说明>"
        url: "https://clawhub.ai/skills/<skill-name>"      # ClawHub
        # 或 SkillHub：
        # url: "https://skillhub.cn/skills/<slug>"
        # 或联想开放平台:
        # url: "https://open.lenovomm.com/skills/<slug>"
        publish_id: "<从命令输出获取>"
        # SkillHub 还要加一行 namespace: "<your-namespace>";ClawHub / 联想无 namespace 字段

Example: Sync skill (publish to the corresponding platform according to the routing policy)

A. MIT tools → ClawHub (using git-batch-commit as an example)

# 1. 检查白名单
grep -A1 "^git-batch-commit:" skills/skill-publish-sync/config/allowlist-clawhub.yaml
# 输出：platforms: [clawhub]   # MIT 工具，只发 ClawHub

# 2. 比较版本
# SKILL.md: version: "1.2.0"
# sync-records.yaml: platforms.clawhub.version: "1.1.0" → 需要同步

# 3. 准备并发布到 ClawHub
bash skills/skill-publish-sync/scripts/prepare-publish.sh skills/git-batch-commit
clawhub publish /tmp/clawhub-publish-git-batch-commit \
  --slug git-batch-commit --name "Git Batch Commit" \
  --version "1.2.0" --changelog "本次变更说明"

# 4. 更新 sync-records.yaml 的 platforms.clawhub 字段

B. CC-BY-NC legal category → SkillHub (using legal-qa-extractor as an example)

# 1. 检查白名单
grep -A2 "^legal-qa-extractor:" skills/skill-publish-sync/config/allowlist-skillhub.yaml
# 输出：platforms: [skillhub]  +  display_name: "法律问答知识提取"

# 2. 准备（--platform skillhub 会自动从配置读 display_name/slug 注入临时副本 frontmatter）
bash skills/skill-publish-sync/scripts/prepare-publish.sh --platform skillhub skills/legal-qa-extractor

# 3. 发布到 SkillHub
skillhub publish /tmp/skillhub-publish-legal-qa-extractor \
  --version "<新版本号>" --changelog "<变更说明>"

# 4. 更新 sync-records.yaml 的 platforms.skillhub 字段

Failure handling

On sync failure, record status: failed
Do not retry; let the user decide the next action
Record the failure reason to changelog_summary

SkillHub-specific notes

Token acquisition

Log in to SkillHub
Generate an API Token in personal settings (format skh_xxx, prefix skh_）
Via skillhub login --key skh_xxx save to local

skillhub login --key skh_xxx
skillhub auth whoami     # 确认身份

Security reminder: A token is equivalent to an account password;absolutely do notwrite it into configuration files or commit it to Git.skillhub login The token has already been saved locally, so subsequent publish commands do not need to pass the token again (the CLI reads it automatically); if you need to temporarily override it, you can use skillhub publish at --token skh_xxx。

slug / namespace and version number rules

Tencent SkillHub uses slug + displayName + namespace(account-bound) to identify skills.slug/displayName are not written into the source SKILL.md, but are configured in config/allowlist-skillhub.yaml(platform-specific whitelist), and before publishing are injected by prepare-publish.sh into the temporary copy's frontmatter:

slug: kebab-case, 2-128 characters (actual validation regex). Default = skill directory name (i.e. name); override with the allowlist-skillhub.yaml field in slug: (to handle duplicate names/names already taken). The slug is unique within the same namespace; duplicate names are allowed across namespaces (e.g. @cat-xierluo/md2word does not conflict with someone else's @xxx/md2word )
namespace: namespace,bound to the account(server-side @<namespace>/<slug> format), the publish commanddoes not need to and cannotspecify it on the command line—the CLI uses the logged-in account's identity, and the server automatically assigns it under your namespace. Your namespace can be determined via skillhub search <你的skill> by looking at the @xxx/slug prefix of a published skill
version: valid SemVer (major.minor.patch), e.g. 1.0.0. It is recommended to start from 0.1.0 or 1.0.0 . Published versions cannot be modified; you can only publish a new version

API host

The default API host is https://api.skillhub.cn. To override it,skillhub publish use --host <URL> to specify it.

SKILL.md frontmatter requirements

When publishing to Tencent SkillHub, skill metadata is read from the SKILL.md frontmatter (of a temporary copy).The source SKILL.md does not need to include slug/displayName——prepare-publish.sh will read from config/allowlist-skillhub.yaml and inject into the temporary copy:

Field	Source	Description
`slug`	Defaults to `name`；`allowlist-skillhub.yaml` of `slug:` can be overridden	kebab-case, 2-128 characters
`displayName`	`allowlist-skillhub.yaml` of `display_name:`(required)	SkillHub display name (Chinese)
`version`	Source SKILL.md frontmatter (or `--version` override)	Valid SemVer
`summary`	Source SKILL.md (optional)	Short summary
`description`	Source SKILL.md (optional)	Longer description
`tags`	Source SKILL.md (optional)	Tag array
`license`	Source SKILL.md (optional)	License identifier
`homepage`	Source SKILL.md (optional)	Homepage URL

Field differences between ClawHub and Tencent: ClawHub uses the name field + command-line --slug/--name; Tencent uses slug + displayName(injected into the temporary copy, not written to the source file; the namespace is bound to the account). The source SKILL.md keeps only standard fields such as name/description/version ; the identification needs of both platforms are satisfied by the publishing process (command-line parameters / config injection), without polluting the skill itself.

CLI version notes

Tencent skillhub CLI is a Python script.Older versions (< 2026.7.29) do not support publish/login, run skillhub self-upgrade to upgrade to the latest version; if unsure, first use skillhub self-upgrade --check-only to check whether a new version is available.

Publishing rate limits (avoid concurrent throttling)

The Tencent SkillHub API has rate limits on publish requests.Multiple consecutive skillhub publish will trigger throttling(error "请求过于频繁,请稍后再试"), a single publish failure needs to be retried, and it will slow down the entire batch process.

Actual measured safe interval: for each skillhub publish between sleep 12-15 seconds. Batch publishing N items estimated total time ≈ N × (single publish time + 15s).

Recommended approach（Sequential + sleep, do not parallelize）:

while IFS='|' read -r s v; do
  bash prepare-publish.sh --platform skillhub "skills/$s" >/dev/null 2>&1
  echo "$s: $(skillhub publish /tmp/skillhub-publish-$s --version "$v" --changelog '...' 2>&1 | tail -1)"
  sleep 15
done <<EOF
skill-a|1.0.0
skill-b|1.1.0
EOF

Failure handling: When encountering a rate limit error, do not immediately rerun the entire batch.sleep 30-60 seconds later, retry the failed skill individually (from sync-records.yaml see which platforms are still pending）。

Avoid：for s in ...; do (skillhub publish $s &); done(parallel) —— 100% triggers rate limiting.

Lenovo Open Platform (LenovoSkill CLI) upload

Lenovo Open Platform (https://open.lenovomm.com) is a skill distribution channel for the Lenovo AI agent ecosystem.Under the v1.7.1 three independent list architecture, Lenovo is the third channel parallel to ClawHub / SkillHub——any skill can be registered in allowlist-lenovo.yaml (no license restrictions), with fields maintained independently from the ClawHub / SkillHub platforms. One-to-many: the same skill can appear simultaneously in allowlist-clawhub.yaml + allowlist-skillhub.yaml + allowlist-lenovo.yaml multiple files.

Differences from the first two platforms

Dimension	ClawHub	Tencent SkillHub	Lenovo Open Platform
Platform role	Independent list platform（`allowlist-clawhub.yaml`）	Independent list platform（`allowlist-skillhub.yaml`）	Independent list platform（`allowlist-lenovo.yaml`, parallel to SkillHub / ClawHub)
Workflow	Single `publish` command	Single `publish` command	Two steps：`package` zip → `push` upload zip
CLI	`clawhub`	`skillhub`（Python）	`lenovoskill`（`npx @lenovo-open/skill-cli`，Node.js）
Credentials	`clawhub login`	`skillhub login --key skh_xxx`	`lenovoskill login`(OAuth browser authorization)
Identifier mechanism	Command line `--slug/--name`	Temporary copy frontmatter (slug/displayName)	Temporary copy frontmatter (slug/displayName) + in-project `.skill-config.json`
namespace	None	Yes(bound to account, not passed in publish command, used for tracing records)	None(OAuth dual token)
License	Mandatory MIT-0	Unrestricted	Unrestricted
Sync record	`sync-records.yaml` `platforms.clawhub`	`sync-records.yaml` `platforms.skillhub`(including namespace)	`sync-records.yaml` `platforms.lenovo`(no namespace)

CLI installation

# 方式 1：npx（推荐，免装、始终最新版）
npx @lenovo-open/skill-cli <cmd>

# 方式 2：全局安装（短命令、可离线）
npm install -g @lenovo-open/skill-cli
# 之后用 lenovoskill 代替 npx @lenovo-open/skill-cli

Hereafter uniformly use lenovoskill; when using npx, just replace it.Prerequisite dependencies: Node.js ≥ 18 + npm(see the "Dependencies · System Dependencies" section of this document).

First login

lenovoskill login          # 浏览器 OAuth 授权
lenovoskill whoami         # 确认身份
lenovoskill logout         # 退出登录

Credentials and configuration (OAuth secure storage,forbidden to write into repository or public platforms）：

File	Mode	Purpose
`~/.lenovoskill/auth.json`	600	Login Token + Biz Token(OAuth dual token) + user information
`~/.lenovoskill/config.json`	Default	API URL (default `https://open.lenovomm.com`) + default visibility

Upload workflow (5 steps)

# ① 准备安全过滤目录（复用 prepare-publish.sh，会触发 SkillHub/联想 共用的 frontmatter 注入段）
bash skills/skill-publish-sync/scripts/prepare-publish.sh --platform lenovo skills/<skill-name>
# → 生成 /tmp/lenovo-publish-<skill-name>;临时副本 SKILL.md 已注入 slug + displayName

cd /tmp/lenovo-publish-<skill-name>

# ② 准备 .skill-config.json（联想 CLI 项目级元数据,与 SKILL.md frontmatter 互补）
#    - 源 skill 目录若已带 .skill-config.json,prepare-publish.sh 已一并复制,直接用
#    - 否则从源目录复制,或在此目录跑一次 lenovoskill init 生成默认配置再按需编辑
lenovoskill init            # 生成默认 .skill-config.json(可手编 ignore patterns)

# ③ 打 zip
lenovoskill package
# → 生成 <skill-name>.zip

# ④ 上传 zip
lenovoskill push

⚠️ Must first pass prepare-publish.sh: Lenovo CLI has no built-in .gitignore filtering; directly packaging the source directory will include local real config/*.yaml、scripts/ intermediate artifacts,.DS_Store and so on into the zip for upload. Run prepare-publish.sh first; the temporary directory is already the "clean content" after git ls-files filtering, so only then is it safe to zip.

`.skill-config.json`

The Lenovo CLI reads this file in the root directory of each skill project, containing:

Skill metadata: name / version / description, etc. (can be consistent with the SKILL.md frontmatter)
Ignore patterns: files to additionally exclude when packaging (double insurance in coordination with prepare-publish.sh)

Recommendation: this fileis only generated/copied in the temporary directory, and is not written back to the source skill repository — it is merely a private adaptation layer of the Lenovo CLI and has nothing to do with the standard skill structure.

Relationship with frontmatter slug/displayName: the Lenovo CLI simultaneously reads from the SKILL.md frontmatter slug + displayName(injected into the temporary copy by prepare-publish.sh) and from the project's .skill-config.json to read metadata.The two should remain consistent——displayName / slug must not conflict between the SKILL.md frontmatter and .skill-config.json .

Command quick reference

Command	Purpose
`lenovoskill login`	Log in to the Lenovo Open Platform (OAuth)
`lenovoskill logout`	Log out
`lenovoskill whoami`	Show the currently logged-in user
`lenovoskill init`	Initialize .skill-config.json in the current directory
`lenovoskill package`	Package the current directory into a zip
`lenovoskill push`	Upload the zip to the Lenovo Open Platform

Failure handling

push fails: first run lenovoskill whoami to confirm the login state; check .skill-config.json fields; confirm the zip is the one package just generated.
Packaged content is incorrect:Return to step 1, modify the source skill and rerun prepare-publish.sh, do not repackage directly in the source directory (this would bypass the security filtering).
Want to revoke a published version: refer to the "Fixing Published Skills" section of this document, but contacting Lenovo platform support is left to the user's own decision and is outside the scope of this skill.

Relationship with sync-records.yaml

Lenovo platform publish resultsmustbe written into sync-records.yaml the platforms.lenovoof — just like SkillHub / ClawHub, it is a fixed component of the three-platform independent whitelist architecture.platforms.lenovo The field is isomorphic with platforms.skillhub / platforms.clawhub (version / last_sync / git_hash / status / changelog_summary / url / publish_id），no namespace field(OAuth dual-token mechanism, with no concept of namespace).

records:
  <skill-name>:
    platforms:
      lenovo:                      # 三平台独立白名单架构(联想是独立第三选项,允许与其他平台字段独立维护)
        version: "<新版本号>"
        last_sync: "<ISO 8601 时间>"
        git_hash: "<当前 commit hash>"
        status: synced
        changelog_summary: "<变更说明>"
        url: "https://open.lenovomm.com/skills/<slug>"
        publish_id: "<从 lenovoskill push 输出获取>"
        # 无 namespace:联想用 OAuth 双 token(Login Token + Biz Token)

⚠️ NOT_VERIFIED: the details of the above LenovoSkill CLI commands (login/init/package/push behavior,.skill-config.json fields, OAuth flow, platform license review,whether slug/displayName are actually read from the SKILL.md frontmatter)come from materials provided by the user and have not been independently tested. For the first actual push, the CLI's actual behavior shall prevail; if the behavior deviates (for example, if testing finds that the Lenovo CLI does not read the frontmatter and relies entirely on .skill-config.json), come back and correct this section — the inject section may be meaningless for lenovo, and this decision needs to be re-evaluated.

Sync strategy

Version number handling

Extract the version number from the CHANGELOG.md first line of the skill
Format requirements:## [x.y.z] - YYYY-MM-DD
Automatically handle the v prefix (v1.0.0 → 1.0.0）

Automatic fields

Field	Handling method
`homepage`	Automatically set to the GitHub repository address
`version`	Extracted from CHANGELOG.md (if not specified in SKILL.md)

Sync scope control (allowlist mechanism, three independent lists in v1.7.1)

Configuration files: Three independent allowlist files (v1.7.1 architecture, see DECISIONS D-2026-08-12-02）：

skills/skill-publish-sync/config/allowlist-clawhub.yaml(ClawHub platform allowlist)
skills/skill-publish-sync/config/allowlist-skillhub.yaml(SkillHub platform allowlist)
skills/skill-publish-sync/config/allowlist-lenovo.yaml(Lenovo Open Platform allowlist)

Priority: allowlist > default ignore rules

If the corresponding platform's allowlist-${PLATFORM}.yaml exists: only sync the skills listed in that file (each platform independently reads its own allowlist file)
If the corresponding platform's allowlist-${PLATFORM}.yaml does not exist: use the default ignore rules (ignore test/, private-skills/, node_modules/)

Configuration format (three independent lists in v1.7.1, each file has no platforms field):

# allowlist-clawhub.yaml（17 条 MIT 通用工具,display_name/slug 可选）
md2word:
patent-download:
# ...

# allowlist-skillhub.yaml（12 条 CC-BY-NC 法律类,display_name 必填）
legal-qa-extractor:
  display_name: "法律问答知识提取"
contract-copilot:
  display_name: "合同起草与审查助手"
# ...

# allowlist-lenovo.yaml（0 条,初始空,用户按需手动添加）
# contract-copilot:
#   display_name: "合同起草与审查助手"    # 与 skillhub 分区独立维护,可不同

One-to-many semantics: the same skill can appear in multiple files at the same time (e.g., contract-copilot can be simultaneously registered in allowlist-skillhub.yaml and allowlist-lenovo.yaml , each maintaining its own independent display_name). There is no need to modify the old skill's configuration for a new platform.

If you need to enable/disable a skill on a certain platform, simply add/comment out that skill's line in the corresponding platform's allowlist file.

Configuration files and privacy (example vs local, three independent lists in v1.7.1)

Configuration files are split into a public template and a local real version, making it easy for this skill to be reused by others:

File	Role	Whether committed	Contains local real data
`config/allowlist-clawhub.example.yaml`	ClawHub allowlist public template	✅ Committed	None
`config/allowlist-skillhub.example.yaml`	SkillHub allowlist public template	✅ Committed	None
`config/allowlist-lenovo.example.yaml`	Lenovo allowlist public template	✅ Committed	None
`config/sync-records.example.yaml`	Sync record public template	✅ Committed	Uses placeholders
`config/allowlist-clawhub.yaml`	ClawHub local real allowlist	❌ Excluded by gitignore	—
`config/allowlist-skillhub.yaml`	SkillHub local real allowlist	❌ Excluded by gitignore	—
`config/allowlist-lenovo.yaml`	Lenovo local real allowlist	❌ Excluded by gitignore	—
`config/sync-records.yaml`	Local real sync record	❌ Excluded by gitignore	Fill in the actual publish_id, etc.

First-time use: copy the corresponding .example.yaml and remove the .example suffix, then fill in your actual data. The three allowlist filesare maintained independently——the same skill can be registered independently in different allowlists (one-to-many). In the SkillHub / Lenovo allowlist files, display_name is required (Chinese display name, injected into the temporary copy's frontmatter by prepare-publish.sh before publishing; the source SKILL.md does not contain this field).sync-records.yaml records publish_id and other publishing results.

The root directory .gitignore 's **/config/*.yaml + !**/config/*.example.yaml rulestake effect for all platforms (ClawHub / SkillHub / Lenovo)(shared prepare-publish.sh filter logic). When actually publishing skill-publish-sync itself, the three real allowlist-*.yaml and sync-records.yaml all do not enter the temporary directory; only the .example.yaml template will be uploaded.

The root directory .gitignore already passes the **/config/*.yaml + !**/config/*.example.yaml rules, automatically excluding real configuration and retaining the template.

File Filtering Rules

When publishing, .gitignore filtering rules are automatically applied to ensure that sensitive files and temporary files are not uploaded. All three platforms share the same filtering logic.

Dual Filtering Mechanism：

Project Root .gitignore - Automatically detect the Git repository root's .gitignore
Skill Internal .gitignore - If the skill directory has its own .gitignore, it will additionally apply

Default exclusions(always in effect):

.git/ - Git directory
node_modules/ - Node.js dependencies
__pycache__/ - Python cache
.DS_Store - macOS system files

Sync Process

Before each sync, the following will automatically:

Create a temporary directory - The prefix varies by platform:/tmp/clawhub-publish-<skill> or /tmp/skillhub-publish-<skill>
Copy the filtered files - Use rsync to copy files following .gitignore rules
Publish to the target platform - Execute the corresponding platform's publish command from the temporary directory
Clean up the temporary directory - Automatically cleaned up after publishing is complete

Manually Prepare the Publish Directory

If you need to manually check the files to be published:

# ClawHub（默认）
bash skills/skill-publish-sync/scripts/prepare-publish.sh skills/trademark-assistant
ls -la /tmp/clawhub-publish-trademark-assistant/

# 腾讯 SkillHub
bash skills/skill-publish-sync/scripts/prepare-publish.sh --platform skillhub skills/trademark-assistant
ls -la /tmp/skillhub-publish-trademark-assistant/

Security Best Practices

Pre-publish Checklist

Confirm that .gitignore includes all sensitive file patterns
Use prepare-publish.sh to check the files to be published
Do not include API keys, passwords, etc. in the skill
Use .env.example instead of .env files

Common Sensitive Files

.env - Environment variables (use .env.example as a template)
config.yaml - Configuration files (use config.example.yaml as a template)
*.db, *.sqlite - Database files
logs/ - Log directories
downloads/, output/ - Output directories

Fix a Published Skill

If you find that a published skill contains sensitive information:

Update immediately - Delete the sensitive files from the skill directory
Update .gitignore - Ensure they will not be included again in the future
Republish - Use clawhub publish / skillhub publish / lenovoskill push to update the corresponding platform
Contact platform support - If the old version needs to be deleted

Important Reminder：

Both platforms are public platforms: anyone can access the published skill
Do not include customer information: case files, communication records, etc. should be excluded
Do not include credentials: API keys, tokens, etc. should use environment variables

FAQ

Version number not updated?

Check the CHANGELOG.md format:

## [1.0.0] - 2026-03-21

### 新增
- 新功能描述

Sync failed?

ClawHub: Run clawhub sync --dry-run to check the configuration
SkillHub: Confirm skillhub auth whoami login is working,config/allowlist-skillhub.yaml configured for this skill display_name(slug defaults to name; if display_name is missing, prepare-publish.sh will fail-closed with an error)
Lenovo Open Platform: Confirm lenovoskill whoami login is working,config/allowlist-lenovo.yaml configured display_name(Lenovo also has frontmatter injected by prepare-publish.sh); Node.js ≥ 18 + npx available; temp directory generated .skill-config.json(generated by init or copied from source)
Confirm that the SKILL.md frontmatter format is correct
Check the allowlist: whether the target skill is in the corresponding platform's allowlist-${PLATFORM}.yaml (v1.7.1 has three independent lists, with a separate file per platform)

ClawHub reporting a license conflict?

Licenses such as CC-BY-NC conflict with ClawHub MIT-0 and cannot be published to ClawHub. Skills of this kind should be published to SkillHub instead (no license restrictions). See ClawHub Skill Format official documentation。

Input/Output

Input

Required: the locally developed skill directory
Optional: specify the skill name list, target platform (clawhub / skillhub / lenovo, with a separate allowlist file per platform allowlist-${PLATFORM}.yaml; see the platform comparison table and the sync scope control section for details), allowlist configuration

Output

Sync result report (success/failure list, grouped by platform)
Error messages (if any)

Sync records

After each sync, config/sync-records.yaml will be updated as a record file to facilitate traceability and incremental sync. Records are stored separately by platform.

Record structure

records:
  <skill-name>:
    platforms:
      clawhub:    # ClawHub 平台记录
        <fields>
      skillhub:   # 腾讯 SkillHub 平台记录(含 namespace 字段)
        <fields>
      lenovo:     # 联想开放平台记录(无 namespace 字段)
        <fields>

Record fields

Field	Description	Applicable platform
`version`	Version number at sync time	All three
`last_sync`	Last sync time (ISO 8601)	All three
`git_hash`	Commit hash at sync time	All three
`status`	`synced` / `pending` / `failed` / `skipped` / `slug_conflict` / `deleted`	All three
`changelog_summary`	Change summary	All three
`url`	Platform publish URL	All three
`publish_id`	Platform internal ID	All three
`namespace`	Namespace (account binding, such as `cat-xierluo`; not passed during publishing, recorded for traceability of install commands)	SkillHub only(ClawHub / Lenovo have no namespace concept)

Record example

records:
  md2word:
    platforms:
      clawhub:
        version: "1.1.8"
        last_sync: "2026-03-24T16:42:00+08:00"
        git_hash: "fbb1db4"
        status: synced
        changelog_summary: "全书自动更新目录域"
        url: "https://clawhub.ai/skills/md2word"
        publish_id: "k97dtn63cty6ezwzj9g9ry85818axkdq"
      skillhub:
        namespace: "cat-xierluo"
        version: "1.1.8"
        last_sync: "2026-08-01T10:00:00+08:00"
        git_hash: "fbb1db4"
        status: synced
        changelog_summary: "首次发布 SkillHub"
        url: "https://skillhub.cn/cat-xierluo/md2word"
        publish_id: "sh_xxxxxxxxxxxxxxxx"
      lenovo:
        # 无 namespace:联想用 OAuth 双 token(Login Token + Biz Token)
        version: "1.1.8"
        last_sync: "2026-08-12T10:00:00+08:00"
        git_hash: "fbb1db4"
        status: synced
        changelog_summary: "首次发布联想开放平台"
        url: "https://open.lenovomm.com/skills/md2word"
        publish_id: "len_xxxxxxxxxxxxxxxx"

Purpose

Incremental sync: only sync status: pending or skills with version updates (judged independently per platform)
Traceability: use git_hash to trace the code state at publish time
Quick access: use url to directly access the skill page on the platform

About this skill

Dependencies

System dependencies

CLI tools (install on demand per target platform)

First-time installation checklist (SkillHub publishing)

Platform comparison

Prerequisites

Required SKILL.md frontmatter fields

CLI Installation

Usage

1. Log in to the platform (first-time use)

2. Verify login status

3. Sync skills

ClawHub

Tencent SkillHub

Single Skill Sync Workflow

Pre-checks

Version detection

Execute sync

Step 1: Prepare the publish directory

Step 2: Execute publishing

Step 3: Update the sync record

Example: Sync skill (publish to the corresponding platform according to the routing policy)

Failure handling

SkillHub-specific notes

Token acquisition

slug / namespace and version number rules

API host

SKILL.md frontmatter requirements

CLI version notes

Publishing rate limits (avoid concurrent throttling)

Lenovo Open Platform (LenovoSkill CLI) upload

Differences from the first two platforms

CLI installation

First login

Upload workflow (5 steps)

.skill-config.json

Command quick reference

Failure handling

Relationship with sync-records.yaml

Sync strategy

Version number handling

Automatic fields

Sync scope control (allowlist mechanism, three independent lists in v1.7.1)

Configuration files and privacy (example vs local, three independent lists in v1.7.1)

File Filtering Rules

Sync Process

Manually Prepare the Publish Directory

Security Best Practices

Pre-publish Checklist

Common Sensitive Files

Fix a Published Skill

FAQ

Version number not updated?

Sync failed?

ClawHub reporting a license conflict?

Input/Output

Input

Output

Sync records

Record structure

Record fields

Record example

Purpose

`.skill-config.json`