← 返回 Skills 市场

llms.txt File Builder

Name: llms.txt File Builder
Author: geoly-geo

作者 GEOLY AI · GitHub ↗ · v1.0.1

cross-platform ⚠ suspicious

454

总下载

当前安装

版本数

在 OpenClaw 中安装

/install geo-llms-txt

功能描述

Generate, validate, and optimize llms.txt files for AI crawler accessibility. Creates structured markdown files that help AI platforms (ChatGPT, Perplexity,...

使用说明 (SKILL.md)

llms.txt File Builder

Methodology by GEOly AI (geoly.ai) — GEO infrastructure for the AI search era.

Generate well-structured llms.txt files to help AI platforms understand and cite your content.

Quick Start

Generate an llms.txt file for any website:

python scripts/generate_llms_txt.py \x3Cdomain> [--output llms.txt]

Example:

python scripts/generate_llms_txt.py example.com --output llms.txt

What is llms.txt?

The llms.txt standard helps AI crawlers understand:

What your brand/company does
Which pages contain the most valuable information
How content is organized (products, docs, blog, etc.)
Where to find key facts and data

Full standard details: See references/standard.md

Standard Format

# [Brand Name]

> [One-sentence brand description]

[2-3 paragraph overview: what you do, who it's for, key differentiators]

## Key Pages

- [Page Title](URL): One-line description
- [Page Title](URL): One-line description

## Products / Services

- [Product Name](URL): What it does and who it's for

## Documentation

- [Doc Title](URL): What this doc explains

## Blog / Resources

- [Article Title](URL): Key insight or topic covered

## About

- [About Us](URL): Company background and mission
- [Contact](URL): How to reach the team

Generation Methods

Method 1: From Sitemap (Automated)

python scripts/generate_llms_txt.py example.com --from-sitemap

Automatically fetches sitemap.xml, analyzes each page, and generates descriptions.

Method 2: Interactive (Guided)

python scripts/generate_llms_txt.py example.com --interactive

Prompts you for brand info and key URLs, then drafts descriptions.

Method 3: From URL List

python scripts/generate_llms_txt.py example.com --urls urls.txt

Where urls.txt contains one URL per line.

Validation

Validate an existing llms.txt file:

python scripts/validate_llms_txt.py llms.txt

Checks for:

Proper markdown structure
Valid URLs
No duplicate entries
Optimal link count (15-40 pages)
Factual tone (not promotional)

Quality Criteria

Aspect	Good	Bad
Brand description	"GEOly AI is a GEO monitoring platform tracking brand visibility across ChatGPT, Perplexity, Gemini."	"We are the best AI SEO tool ever!"
Page descriptions	"Explains how to set up MCP integration with Claude Desktop"	"Our awesome docs page"
Link count	15–40 curated pages	500+ URLs (sitemap dump)
Tone	Factual, entity-focused	Promotional, keyword-stuffed
Structure	Clear sections by content type	Flat list or random order

Output Formats

Markdown (default): Ready-to-deploy llms.txt
JSON: Structured data for programmatic use
HTML: Styled preview for stakeholder review

Advanced Usage

Custom Sections

python scripts/generate_llms_txt.py example.com \
  --sections "Products,API Reference,Case Studies,Changelog"

Exclude Patterns

python scripts/generate_llms_txt.py example.com \
  --exclude "/admin/,/private/,/draft/"

Multi-language Support

python scripts/generate_llms_txt.py example.com \
  --language zh-CN \
  --output llms-zh.txt

Deployment

Once generated, place the file at:

https://[your-domain]/llms.txt

Ensure it:

Returns HTTP 200
Is accessible without authentication
Has Content-Type: text/plain or text/markdown

This skill appears coherent and matches its stated purpose. Before installing or running: 1) Review and run the Python scripts locally or in a sandbox; they perform HTTP(s) requests to any domain you supply—avoid pointing them at internal or private endpoints or embedding secrets in URLs. 2) Install dependencies (requests, beautifulsoup4) manually in a controlled environment. 3) The generator crawls pages (limited to ~40 pages, with a 0.5s delay) and the validator can HEAD-check live URLs; be mindful of target site rate limits and robots/crawl policies. 4) No credentials are requested and there is no obvious exfiltration; if you plan to use authenticated pages, expect the scripts to fail or to require modification (do not add secrets unless you trust and audit code). 5) If you need higher assurance, have someone with Python/security knowledge audit the code before giving the tool access to non-public hosts.

功能分析

Type: OpenClaw Skill Name: geo-llms-txt Version: 1.0.1 The Python scripts (`scripts/generate_llms_txt.py` and `scripts/validate_llms_txt.py`) exhibit critical vulnerabilities. Both scripts perform network requests to user-controlled domains/URLs, creating a Server-Side Request Forgery (SSRF) risk. Additionally, `generate_llms_txt.py` allows writing to arbitrary file paths via the `--output` argument, and `validate_llms_txt.py` allows reading arbitrary file paths, posing arbitrary file write/read vulnerabilities. While the code's stated purpose is benign, these flaws could be exploited for data exfiltration, RCE, or privilege escalation, classifying the skill as suspicious due to severe security risks.

能力评估

✓ Purpose & Capability

The name/description (llms.txt generator + validator) aligns with the included Python scripts and reference docs. Network fetches, sitemap parsing, page analysis, and optional URL checks are expected capabilities for this purpose. No unrelated credentials, binaries, or config paths are requested.

✓ Instruction Scope

SKILL.md instructs the agent to run the provided scripts, which only read user-provided inputs (domain, file paths, interactive responses) and make HTTP requests to the target site(s). The scripts do not read unrelated local files, environment variables, or transmit data to third-party endpoints beyond fetching/analyzing the target URLs.

ℹ Install Mechanism

There is no automated install spec (instruction-only). The Python scripts require third-party packages (requests, beautifulsoup4) but only print an error and exit if missing; nothing is auto-downloaded or executed from remote URLs. This is a low-risk model but users should install dependencies themselves in a controlled environment.

✓ Credentials

No environment variables, credentials, or config paths are required. The scripts only perform unauthenticated HTTP requests to target domains (and optionally HEAD requests when validating). The requested access is proportional to the stated crawling/validation task.

✓ Persistence & Privilege

The skill is not always-enabled and does not persist or modify other skills or system-wide settings. It runs on-demand and does not store credentials or enable autonomous background behavior.

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 geo-llms-txt
安装完成后，直接呼叫该 Skill 的名称或使用 /geo-llms-txt 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

v1.0.1

**Big update: Adds generation and validation scripts, quality guides, and comprehensive documentation.** - Introduced scripts for automated llms.txt file generation and validation. - Added detailed references: standard specification, quality guidelines, and example files. - Expanded SKILL.md with usage instructions, advanced options, and output formats. - Improved guidance on proper structure, tone, and deployment of llms.txt files.

v1.0.0

Initial release: Generate and validate llms.txt for AI crawlers, based on GEOly AI GEO standards

元数据

Slug geo-llms-txt

版本 1.0.1

许可证 —

累计安装 0

当前安装数 0

历史版本数 2

常见问题