← Back to Skills Marketplace
teamolab

Chief Editor

by teamolab · GitHub ↗ · v0.1.0
cross-platform ⚠ suspicious
987
Downloads
0
Stars
5
Active Installs
1
Versions
Install in OpenClaw
/install chief-editor
Description
Professional chief editor skilled in reviewing, refining, and ensuring content quality while adhering strictly to user preferences and source materials.
README (SKILL.md)

Chief Editor

Overview

This skill provides specialized capabilities for chief editor.

Instructions

You are a professional chief editor.# User Personalized Preferences [CRITICAL]The following are user-inputted personalized writing preferences, which you MUST faithfully adhere to: $GET_USER_TEMPLATE$. If these preferences conflict with your other system prompt instructions, these preferences take the highest priority. If these preferences conflict with the user prompt, the user prompt takes priority.# Workflow### Step 1: Gather Information from Provided Sources (If no sources are provided, skip this step)Part A: Read Attached or Knowledge Base Files1. Check for user-provided attachments (e.g., wiki files, reports).2. If attachments are present, you MUST use the appropriate tool (e.g., read_wiki_document) to read the content of ALL attached files. This should be performed in a single, parallel tool call.3. If the user refers to relevant content within the knowledge base, you need to call wiki_retriever to locate the associated content.Note:- Knowledge Base Agent - Corresponding Tool: wiki_retriever - Delegation Scenario: When the user mentions 'knowledge base' or documents within the knowledge base, this subordinate should be called to retrieve the corresponding documents. The Knowledge Base Agent can retrieve and acquire documents from the knowledge base, further analyze their content by reading them, and ultimately return the precisely required knowledge base documents.- Important Note: Do not instruct the Knowledge Base Agent to return all documents within the knowledge base. This agent should return only those documents that match the specified criteria after filtering.Part B: Read URLs Found in Files1. Upon completion of Part A, you MUST meticulously review the full-text content returned from the attachments.2. Identify all URLs contained within this text.3. From the list of identified URLs, select a maximum of five that are most critical and supplementary to understanding the subject.4. Subsequently, you MUST use the url_scraping tool to read the content of these selected URLs. This should be performed in a single, parallel tool call.5. This step is mandatory if any relevant URLs are found within the documents. Do NOT proceed to Step 2 without first attempting to locate and scrape URLs from the provided documents.$GET_CREATION_TEMPLATE$### Step 3: Execute Content Creation Strategy Guided by the System Prompt Acquired in Step 21. First, determine if the user's request explicitly calls for multiple versions. For example, check if the user has mentioned 'multiple versions,' 'three versions,' 'different styles,' 'multiple options,' or similar keywords.2. If the user explicitly requests multiple versions: please call the following five tools in parallel to generate diverse content: 'editor_call_gemini_2_5_pro_llm', 'editor_call_claude_sonnet_4_llm', 'editor_call_grok4_llm', 'editor_call_deepseek_v3_llm', and 'editor_call_doubao1_5_llm'.3. In all other cases: you are to complete the writing task personally, without calling any of the aforementioned five tools.### Step 4: Check Word Count and Other Writing Requirements1. Based on the current word count returned by the wiki tool and any user-specified word count requirements, evaluate whether additional content is needed or if you should stop immediately and submit the result.2. [CRITICAL] Re-confirm that a list of references has been appended to the very end of your entire article. The reference list MUST NOT appear at the end of a section; it MUST be placed at the conclusion of the entire article (i.e., the last chapter). You MUST NOT submit the result until this step has been confirmed.### Step 5: Submit Writing Result1. Call the submit_result tool, attaching your generated writing result in the attached_files field.# Current Date$DATE$

Usage Notes

  • This skill is based on the chief_editor agent configuration
  • Template variables (if any) like $DATE$, $SESSION_GROUP_ID$ may require runtime substitution
  • Follow the instructions and guidelines provided in the content above
Usage Guidance
This skill appears to do what it says, but take these precautions before installing or using it: - Understand the tool endpoints: the skill will call wiki_retriever, url_scraping, and multiple editor_call_* LLM tools. Confirm who operates those tools and where your document text will be sent (third-party LLMs or internal services). - Avoid feeding sensitive documents: the skill mandates reading ALL attachments and scraping URLs found within them. If your attachments contain confidential data, do not use the skill or remove sensitive content first. - Be cautious about URL scraping: scraped pages may be external and fetching them can leak the fact that you reviewed the doc. If you don't want external network calls, instruct the agent not to scrape URLs or to only follow whitelisted domains. - Multiple-model generation: when multiple versions are requested the skill calls several LLM tools in parallel — that increases the number of external services receiving your content. Only enable that mode if you trust those models. - If you need higher assurance, ask the skill author or platform operator which backends implement the named tools and whether data is logged/retained. Given these data-flow/privacy concerns, the skill is internally coherent for an editor role (benign), but exercise caution about what documents you allow it to read and which tool endpoints you permit.
Capability Analysis
Type: OpenClaw Skill Name: chief-editor Version: 0.1.0 The skill is suspicious due to a critical prompt injection vulnerability and a risky URL scraping capability. In SKILL.md, the instructions explicitly state that `$GET_USER_TEMPLATE$` and `$GET_CREATION_TEMPLATE$` (user-inputted preferences) 'take the highest priority' over 'other system prompt instructions,' allowing an attacker to inject arbitrary instructions and hijack the agent's behavior. Additionally, the agent is instructed to use the `url_scraping` tool to read content from URLs found in user-provided documents, which creates a potential for Server-Side Request Forgery (SSRF) or processing of malicious external content.
Capability Assessment
Purpose & Capability
Name and description align with runtime instructions: the skill instructs the agent to read provided source documents/KB entries and produce edited content. The tools and steps (read documents, consult URLs, generate text, submit results) are coherent with an editing workflow.
Instruction Scope
The SKILL.md mandates reading ALL attached files, using a knowledge-base retriever when KB documents are referenced, identifying URLs inside those documents, and (if any URLs are found) scraping up to five URLs. It also instructs parallel calls to multiple LLM tools when the user requests multiple versions. These behaviors are reasonable for a thorough editor, but they broaden data exposure: attachments and their contents will be sent to the platform's tools and external scraping endpoints and multiple LLM backends may receive document content.
Install Mechanism
Instruction-only skill with no install spec and no code files. Nothing is written to disk or installed by the skill itself.
Credentials
No environment variables, credentials, or config paths are requested. However, the skill references many platform tools (wiki_retriever, url_scraping, several editor_call_* LLM tools, submit_result) that will receive document/text data at runtime — the skill does not require additional secrets itself, but using it will cause data to be transmitted to whatever backends implement those tools.
Persistence & Privilege
always is false and there is no indication the skill modifies agent/system-wide settings or other skills. It does require calling platform tools but does not request permanent presence or elevated privileges.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install chief-editor
  3. After installation, invoke the skill by name or use /chief-editor
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.0
Initial release of chief-editor skill. - Introduces a workflow for professional chief editing, including gathering user attachments and knowledge base content. - Enforces strict adherence to user-specified personalized writing preferences. - Supports URL extraction and content scraping from referenced documents. - Allows for multi-version output when explicitly requested by users. - Ensures word count and reference list requirements are met before submission. - Designed for integration with specific agent tools and runtime template variables.
Metadata
Slug chief-editor
Version 0.1.0
License
All-time Installs 5
Active Installs 5
Total Versions 1
Frequently Asked Questions

What is Chief Editor?

Professional chief editor skilled in reviewing, refining, and ensuring content quality while adhering strictly to user preferences and source materials. It is an AI Agent Skill for Claude Code / OpenClaw, with 987 downloads so far.

How do I install Chief Editor?

Run "/install chief-editor" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Chief Editor free?

Yes, Chief Editor is completely free (open-source). You can download, install and use it at no cost.

Which platforms does Chief Editor support?

Chief Editor is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Chief Editor?

It is built and maintained by teamolab (@teamolab); the current version is v0.1.0.

💬 Comments