← 返回 Skills 市场
yonghyeokrhee

LinkedIn Post

作者 Kanto(칸토) · GitHub ↗ · v1.0.1 · MIT-0
cross-platform ⚠ suspicious
70
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install linkedin-post-openclaw-browser
功能描述
Draft, prepare, and publish LinkedIn feed posts through OpenClaw browser automation. Use when a user wants to turn approved post copy into a real LinkedIn fe...
使用说明 (SKILL.md)

linkedin-post

Use this skill for LinkedIn feed posts.

Prerequisites:

  • OpenClaw CLI is installed
  • OpenClaw browser automation is enabled
  • the chosen browser profile is already logged into LinkedIn
  • a gateway token is available either through OPENCLAW_GATEWAY_TOKEN or an openclaw.json config file

Keep the workflow strict:

  1. Confirm the user wants a feed post, not an article.
  2. Finalize the post copy first.
  3. Prepare the composer and fill the content.
  4. Only click Post after explicit user approval.

Workflow

1. Finalize the copy

Do not open the composer until the copy is settled enough to paste.

If the user is still iterating on tone or structure, finish that first.

2. Prepare the browser session

Use OpenClaw browser CLI, not manual shell hacks.

The environment may need the gateway token even when browser support is enabled.

The helper script reads the token in this order:

  1. OPENCLAW_GATEWAY_TOKEN
  2. the file passed with --config
  3. ~/.openclaw/openclaw.json

If browser commands fail unexpectedly, check:

  • the browser profile is correct and already logged into LinkedIn
  • a gateway token is present
  • openclaw browser status --profile \x3Cprofile> succeeds

3. Use the helper script

Use scripts/linkedin_post.py for the normal workflow. It:

  • starts the browser profile if needed
  • opens the LinkedIn feed share modal
  • finds the composer textbox from the snapshot
  • fills the content
  • optionally clicks Post

Prepare only

python3 scripts/linkedin_post.py --content-file /tmp/linkedin_post.txt

or

cat /tmp/linkedin_post.txt | python3 scripts/linkedin_post.py

If the token is not already exported, pass a config file explicitly:

python3 scripts/linkedin_post.py --content-file /tmp/linkedin_post.txt --config ~/.openclaw/openclaw.json

This fills the post body and stops before publishing.

Publish after approval

python3 scripts/linkedin_post.py --content-file /tmp/linkedin_post.txt --publish

Only run --publish after the user clearly says to post it.

Safety rule

Preparing the composer is reversible. Publishing is external side effect.

Before --publish, make sure the user has already approved:

  • the exact body text
  • any link included in the post
  • that this should go to LinkedIn now

If approval is ambiguous, stop at prepare-only mode and ask.

Notes

  • This skill is for https://www.linkedin.com/feed/?shareActive=true, not LinkedIn article publishing.
  • LinkedIn article pages such as /post/new/ or /article/new/ are different flows. Do not confuse them with feed posting.
  • The skill depends on the openclaw CLI binary and browser subcommands being available on the host.
  • The skill assumes LinkedIn authentication already exists in the selected browser profile. It does not perform account login or two-factor setup.
  • After filling the body, re-snapshot if needed to confirm link unfurl and that the Post button is enabled.
  • If refs change, re-run snapshot and let the helper script rediscover them instead of hardcoding stale refs.
安全使用建议
This skill appears to implement exactly what it claims (automating OpenClaw to prepare/publish LinkedIn feed posts), but metadata omissions are concerning. Before installing: 1) Confirm the openclaw CLI is installed on the host and that you trust it. 2) Inspect your ~/.openclaw/openclaw.json (or the file you will pass with --config) to know what tokens the script will read, and consider exporting OPENCLAW_GATEWAY_TOKEN explicitly instead of leaving unexpected tokens in config files. 3) Note the SKILL.md references scripts/linkedin_post.py but the included file is linkedin_post.py — verify the script path before running. 4) Test the helper with prepare-only (no --publish) to validate behavior, and never run with --publish unless you explicitly approve the final text. 5) Ask the publisher/registry to correct the metadata (declare required binaries and OPENCLAW_GATEWAY_TOKEN) so the requirements are transparent. If you cannot verify these points or do not trust the publisher, do not install or grant the skill access to your browser profile and gateway token.
功能分析
Type: OpenClaw Skill Name: linkedin-post-openclaw-browser Version: 1.0.1 The skill bundle provides a legitimate automation workflow for drafting and publishing LinkedIn posts using the OpenClaw browser CLI. The Python script (linkedin_post.py) interacts with the browser to fill text fields and click buttons based on snapshots, while the instructions (SKILL.md) explicitly mandate user approval before any publishing action occurs.
能力评估
Purpose & Capability
The skill is clearly intended to drive the OpenClaw browser CLI to prepare/publish LinkedIn feed posts, which matches the included python helper. However the registry metadata claims no required env vars or binaries while both the SKILL.md and the code require the 'openclaw' CLI and a gateway token (OPENCLAW_GATEWAY_TOKEN or an openclaw.json). That metadata omission is an incoherence.
Instruction Scope
SKILL.md describes a narrow prepare-then-post flow and the python script implements only those steps (open/start/snapshot/fill/click via OpenClaw). The instructions do not request reading unrelated files or exfiltrating data to external endpoints beyond the OpenClaw gateway flow. They correctly warn to avoid publishing without explicit approval.
Install Mechanism
No install spec is present (instruction-only plus a helper script). Nothing is downloaded from external URLs and no archives are extracted. The highest-risk install pattern is not present.
Credentials
The script legitimately needs a gateway token and the openclaw CLI, and it reads ~/.openclaw/openclaw.json or a user-provided config. Those requirements are proportionate to the purpose, but the registry does not declare them. Additionally, reading the user's openclaw.json can expose any tokens contained therein — this is expected for the task but should be clearly declared in metadata so users can decide.
Persistence & Privilege
The skill does not request always:true and does not modify other skills or system-wide settings. It runs as a one-off helper invoked by the agent; autonomous invocation is allowed (platform default) but not combined with other high-risk indicators here.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install linkedin-post-openclaw-browser
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /linkedin-post-openclaw-browser 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.1
Clarify prerequisites and credential handling. Remove user-specific token path examples from the documentation and tighten runtime expectations for safer installation review.
v1.0.0
Initial release. Automates LinkedIn feed post preparation and publishing through OpenClaw browser CLI with an explicit approval gate before posting.
元数据
Slug linkedin-post-openclaw-browser
版本 1.0.1
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 2
常见问题

LinkedIn Post 是什么?

Draft, prepare, and publish LinkedIn feed posts through OpenClaw browser automation. Use when a user wants to turn approved post copy into a real LinkedIn fe... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 70 次。

如何安装 LinkedIn Post?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install linkedin-post-openclaw-browser」即可一键安装,无需额外配置。

LinkedIn Post 是免费的吗?

是的,LinkedIn Post 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

LinkedIn Post 支持哪些平台?

LinkedIn Post 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 LinkedIn Post?

由 Kanto(칸토)(@yonghyeokrhee)开发并维护,当前版本 v1.0.1。

💬 留言讨论