← 返回 Skills 市场
larsderidder

Agent Content Pipeline

作者 larsderidder · GitHub ↗ · v0.2.3
cross-platform ⚠ suspicious
3275
总下载
5
收藏
9
当前安装
1
版本数
在 OpenClaw 中安装
/install agent-content-pipeline
功能描述
Safe content workflow (drafts/reviewed/revised/approved/posted) with human-in-the-loop approval, plus CLI to list/move/review and post to LinkedIn/X. Use when setting up a content pipeline, drafting content, managing review threads, or posting approved content.
使用说明 (SKILL.md)

Content Pipeline Skill

Safe content automation with human-in-the-loop approval. Draft → Review → Approve → Post.

Setup

npm install -g agent-content-pipeline
content init . # Creates folders + global config (in current directory)

For cryptographic approval signatures (password-protected):

content init . --secure

This creates:

  • drafts/ — work in progress (one post per file)
  • reviewed/ — human reviewed, awaiting your revision
  • revised/ — you revised, ready for another look
  • approved/ — human-approved, ready to post
  • posted/ — archive after posting
  • templates/ — review and customize before use
  • .content-pipeline/threads/ — feedback thread logs (not posted)

Your Permissions

Can do:

  • Write to drafts/
  • Read all content directories
  • Revise drafts based on feedback
  • Move revised files to revised/
  • Run content list to see pending content

Cannot do:

  • Move files to approved/ (only the human can approve)
  • Post content
  • Set status: approved

Creating Content

One post per file. Each suggestion or draft should be a single post, not a collection.

File naming: YYYY-MM-DD-\x3Cplatform>-\x3Cslug>.md

Use frontmatter:

---
platform: linkedin    # linkedin | x | reddit (experimental)
title: Optional Title
status: draft
subreddit: programming  # Required for Reddit
---

Your content here.

Tell the human: "Draft ready for review: content review \x3Cfilename>"

The Review Loop

drafts/ → reviewed/ → revised/ → approved/ → posted/
              ↑          │
              └──────────┘
               more feedback
  1. You write draft to drafts/
  2. Human runs content review \x3Cfile>:
    • With feedback → file moves to reviewed/, you get notified
    • No feedback → human is asked "Approve?" → moves to approved/
  3. If feedback: you revise and move to revised/
  4. Human reviews from revised/:
    • More feedback → back to reviewed/
    • Approve → moves to approved/
  5. Posting happens manually via content post

After Receiving Feedback

When you get review feedback:

  1. Read the file from reviewed/
  2. Apply the feedback
  3. Move the file to revised/
  4. Confirm what you changed
  5. (Optional) Add a note: content thread \x3Cfile> --from agent

Platform Guidelines

LinkedIn

  • Professional but human
  • Idiomatic language (Dutch for NL audiences, don't be stiff)
  • 1-3 paragraphs ideal
  • End with question or CTA
  • 3-5 hashtags at end

X (Twitter)

  • 280 chars per tweet (unless paid account)
  • Punchy, direct
  • 1-2 hashtags max
  • Use threads sparingly
  • If Firefox auth fails, you can paste auth_token and ct0 manually

Manual cookie steps:

  1. Open x.com and log in
  2. Open DevTools → Application/Storage → Cookies → https://x.com
  3. Copy auth_token and ct0

Reddit (experimental)

  • Treat as experimental; API and subreddit rules can change
  • Requires subreddit: in frontmatter
  • Title comes from frontmatter title: (or first line if missing)
  • Match each subreddit's rules and tone

Commands Reference

content list                    # Show drafts and approved
content review \x3Cfile>           # Review: feedback OR approve
content mv \x3Cdest> \x3Cfile>        # Move file to drafts/reviewed/revised/approved/posted
content edit \x3Cfile>             # Open in editor ($EDITOR or code)
content post \x3Cfile>             # Post (prompts for confirmation)
content post \x3Cfile> --dry-run   # Preview without posting
content thread \x3Cfile>           # Add a note to the feedback thread

Security Model

The security model separates drafting (AI) from approval/posting (human):

  • ✅ Agent drafts content
  • ✅ Agent revises based on feedback
  • ❌ Agent cannot approve (human approves via content review)
  • ❌ Agent cannot post

Posting is handled manually via CLI — never by the agent directly.

Platform-specific security

Platform Auth Storage Encrypted? Password Required?
LinkedIn Browser profile ✅ Yes ✅ Yes
X/Twitter Firefox tokens ✅ Yes ✅ Yes

Both platforms require password to post. Tokens are extracted from Firefox and encrypted locally.

安全使用建议
This skill is plausible for managing drafts and reviews, but proceed cautiously. Key things to consider before installing or using it: - The skill asks you to globally install an npm package you cannot audit from the bundle; inspect the package source (npm page / GitHub repo) and review its code before running npm install -g. - The SKILL.md tells you to extract browser cookies (auth_token, ct0) manually — avoid copying/pasting session cookies unless you fully trust and have audited the tooling. Prefer official OAuth or API token flows when possible. - Confirm that the CLI will not post automatically or store tokens unencrypted; test with dry-run modes and minimal privileges. - Verify the package homepage/repository (skill.json references a GitHub page) matches the registry listing and that the maintainer is trusted. - If you cannot audit the npm package, consider running it in an isolated environment (VM/container) or decline installing. If you install, restrict the agent's ability to invoke the CLI autonomously (require explicit user invocation) and never provide browser session cookies to the tool.
功能分析
Type: OpenClaw Skill Name: agent-content-pipeline Version: 0.2.3 The skill defines a content pipeline with explicit human-in-the-loop approval. The `SKILL.md` clearly states that the AI agent is restricted from approving or posting content, which are critical security controls. Instructions for handling sensitive tokens (like X/Twitter `auth_token` and `ct0`) are directed at the human user for setting up the `content` CLI tool, not for the agent to extract or misuse. The skill installs its own `agent-content-pipeline` npm package, which is a standard mechanism for CLI-based OpenClaw skills. No evidence of data exfiltration, malicious execution, persistence, or prompt injection attempts to bypass agent safety was found.
能力评估
Purpose & Capability
The name/description (content drafting → review → approve → post) align with the included CLI workflow and commands. However the registry metadata shown to you said 'no install spec' and 'no required binaries', while the included skill.json declares an npm install that provides 'content'/'content-pipeline' CLI binaries — an internal inconsistency. Installing a global npm package to provide the CLI is reasonable for this purpose, but the skill bundle does not provide the package source for local review (the SKILL.md instructs 'npm install -g agent-content-pipeline').
Instruction Scope
The runtime instructions mostly confine the agent to drafting, revising, and moving files and explicitly say the agent 'cannot approve' or 'cannot post'. But the SKILL.md also documents the 'content post' command and gives manual steps for extracting cookies (auth_token and ct0) from Firefox DevTools and pasting them — instructions that involve sensitive credentials/cookies. That instruction expands the scope of what a human might be asked to supply and could enable posting or token misuse if combined with unattended CLI execution. The guidance to paste browser tokens is a risky operation and is not strictly necessary for drafting/reviewing.
Install Mechanism
An npm install (-g agent-content-pipeline) is the declared install mechanism in skill.json and the SKILL.md. Installing a third-party global npm package is a typical way to get a CLI, but it has moderate risk because arbitrary code will be written to the host and run. No package tarball or local code is included in the skill bundle for review, so you cannot audit what the npm package does before installing.
Credentials
The skill declares no required env vars or config paths (and the top-level metadata reported 'none'), which is consistent with the idea of an offline draft/review pipeline. However the instructions for posting to X ask the user to extract and paste auth_token and ct0 cookies — sensitive tokens not declared in requires.env. This is disproportionate to the core drafting/review purpose and creates a potential vector for credential exposure or accidental agent-driven posting if the CLI or agent later uses those tokens.
Persistence & Privilege
always is false and the skill does not request persistent privileges in the manifest. The SKILL.md emphasizes that approval and posting are human actions. There is no evidence the skill attempts to modify other skills or system-wide config. However, because the skill requires installing an npm package, that package will persist on the system and could run code; this is expected but worth auditing.
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 agent-content-pipeline
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /agent-content-pipeline 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
v0.2.3
Initial publish: content pipeline skill wrapper for agent-content-pipeline CLI.
元数据
Slug agent-content-pipeline
版本 0.2.3
许可证
累计安装 10
当前安装数 9
历史版本数 1
常见问题

Agent Content Pipeline 是什么?

Safe content workflow (drafts/reviewed/revised/approved/posted) with human-in-the-loop approval, plus CLI to list/move/review and post to LinkedIn/X. Use when setting up a content pipeline, drafting content, managing review threads, or posting approved content. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 3275 次。

如何安装 Agent Content Pipeline?

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

Agent Content Pipeline 是免费的吗?

是的,Agent Content Pipeline 完全免费(开源免费),可自由下载、安装和使用。

Agent Content Pipeline 支持哪些平台?

Agent Content Pipeline 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 Agent Content Pipeline?

由 larsderidder(@larsderidder)开发并维护,当前版本 v0.2.3。

💬 留言讨论