← 返回 Skills 市场
storyagent_linkedin
作者
Sudhanshu Sharma
· GitHub ↗
· v1.0.0
1188
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install unipile-linkedin
功能描述
Interact with LinkedIn via Unipile API - send messages, view profiles, manage connections, create posts, react to content. Use when the user asks to message someone on LinkedIn, check LinkedIn messages, view LinkedIn profiles, send connection requests, create LinkedIn posts, or interact with LinkedIn content.
安全使用建议
Before installing or running this skill: (1) Recognize it requires UNIPILE_DSN and UNIPILE_ACCESS_TOKEN (the registry metadata fails to declare them) — treat those as sensitive credentials that grant API access to act on your LinkedIn accounts. (2) Verify the skill source/author and whether you trust the Unipile service and the unipile-node-sdk npm package (supply-chain risk). (3) Prefer creating a least-privileged or revocable token on Unipile (short-lived or scoped) rather than using a long-lived master token. (4) Avoid placing tokens in shared workspace files; store them in a secure secret store or ephemeral environment. (5) If you plan to run the CLI, run npm install in an isolated environment and inspect installed packages. (6) If uncertain about provenance or if the metadata mismatch concerns you, treat installation as risky and either ask the publisher for corrected metadata or run the tool in a sandboxed environment.
功能分析
Type: OpenClaw Skill
Name: unipile-linkedin
Version: 1.0.0
The skill is designed to interact with LinkedIn via the Unipile API, as stated in `SKILL.md` and `README.md`. The core script, `scripts/linkedin.mjs`, uses the `unipile-node-sdk` to perform all LinkedIn operations and retrieves API credentials (`UNIPILE_DSN`, `UNIPILE_ACCESS_TOKEN`) from environment variables, which is standard practice. There is no evidence of prompt injection, unauthorized file system access, arbitrary command execution, or data exfiltration to unapproved endpoints. The `resources/last_50_posts.json` file contains only public sample data. All observed behaviors are aligned with the stated purpose.
能力评估
Purpose & Capability
Name/description, README, SKILL.md and the script all consistently describe interacting with LinkedIn via the Unipile API. However the registry metadata reports "Required env vars: none" and "Primary credential: none", while both SKILL.md and scripts require UNIPILE_DSN and UNIPILE_ACCESS_TOKEN. That mismatch is an incoherence: a LinkedIn integration legitimately needs those credentials, but the published metadata fails to declare them.
Instruction Scope
SKILL.md instructs an agent to run the provided CLI script and to set UNIPILE_DSN and UNIPILE_ACCESS_TOKEN (the script reads only these two env vars). The instructions do not reference unrelated files, system secrets, or unexpected external endpoints — all network calls go to the Unipile DSN provided by the user. Note: SKILL.md suggests storing credentials in a workspace file (~/.openclaw/workspace/TOOLS.md), which could lead to broader credential exposure if that workspace is shared or backed up.
Install Mechanism
There is no install spec (instruction-only skill), which is lower-risk. However the repository includes package.json and package-lock.json (dependencies: dotenv, unipile-node-sdk). If a user or agent attempts to run the script, npm install would be needed to satisfy dependencies — that pulls packages from the public npm registry (supply-chain risk). No download-from-arbitrary-URL behavior or obfuscated installers are present.
Credentials
The functionality legitimately requires Unipile credentials (UNIPILE_DSN and UNIPILE_ACCESS_TOKEN). The problem is that these credentials are not declared in the skill's registry metadata, so a user may not be made aware when granting the skill access. These credentials allow API access to Unipile which can act on LinkedIn on the user's behalf — they are high-value secrets and should be explicitly declared and scoped.
Persistence & Privilege
The skill does not request always:true and does not modify other skills or system-wide settings. It is user-invocable and allows autonomous invocation (platform default), which is expected for tools. No elevated persistence or cross-skill privileges are requested.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install unipile-linkedin - 安装完成后,直接呼叫该 Skill 的名称或使用
/unipile-linkedin触发 - 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
v1.0.0
Initial release: LinkedIn integration via Unipile API
- Send and receive LinkedIn messages, including chat management and unread filtering
- View your own or others’ LinkedIn profiles and company profiles with selectable profile sections
- Manage connections: send and cancel invitations, view pending invites, and browse your connections
- Create and get LinkedIn posts, comment, and react with multiple reaction types
- List contacts and chat attendees
- Includes CLI-based usage examples and setup instructions
元数据
常见问题
storyagent_linkedin 是什么?
Interact with LinkedIn via Unipile API - send messages, view profiles, manage connections, create posts, react to content. Use when the user asks to message someone on LinkedIn, check LinkedIn messages, view LinkedIn profiles, send connection requests, create LinkedIn posts, or interact with LinkedIn content. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 1188 次。
如何安装 storyagent_linkedin?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install unipile-linkedin」即可一键安装,无需额外配置。
storyagent_linkedin 是免费的吗?
是的,storyagent_linkedin 完全免费(开源免费),可自由下载、安装和使用。
storyagent_linkedin 支持哪些平台?
storyagent_linkedin 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 storyagent_linkedin?
由 Sudhanshu Sharma(@sudhanshu746)开发并维护,当前版本 v1.0.0。
推荐 Skills