Veille for Openclaw

A modular RSS aggregator for AI agents that fetches, deduplicates, and scores news articles using LLMs.

romain-grosos
v1.2.5
Mar 3, 2026
2
526
3

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install veille

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install veille using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Veille?

Veille provides a sophisticated engine for RSS feed management, designed to serve as a primary intelligence layer for Openclaw Skills. It handles the heavy lifting of fetching articles from over 20 configured sources while implementing robust deduplication strategies. By combining simple URL tracking with Jaccard similarity and named entity recognition, it ensures that your digests remain concise and free of redundant information.

This skill is built for developers who need high-quality data inputs for their AI agents without the bloat of external dependencies. It seamlessly bridges the gap between raw RSS data and actionable insights through an optional LLM scoring layer, allowing users to rank content based on custom priority thresholds. Whether you are building a cybersecurity monitor or a general tech news tracker, Veille offers the security and flexibility required for professional workflows.

Veille Use Cases

  • Automating morning news digests for technical teams.
  • Supplying AI agents with filtered, high-relevance search and research data.
  • Managing high-volume RSS feeds with automated topic deduplication.
  • Archiving daily tech trends to cloud storage like Nextcloud.

How Veille Works

  1. Fetching: The engine collects data from a curated list of RSS feeds based on a configurable lookback window.
  2. Deduplication: Articles are filtered against a 14-day URL history and cross-referenced for topic similarity.
  3. Scoring: Metadata is processed through an OpenAI-compatible API to assign relevance scores and identify top-tier picks.
  4. Dispatching: The final structured JSON is delivered to Telegram, Email, or local files for immediate use or storage.

Veille Setup

To get started with this skill in the Openclaw Skills framework, run the following commands:

# Navigate to the skill directory and run setup
python3 scripts/setup.py

# Initialize the data directories and config
python3 scripts/init.py

# Execute a full pipeline to fetch, score, and send news
python3 scripts/veille.py fetch --filter-seen --filter-topic | python3 scripts/veille.py score | python3 scripts/veille.py send

Veille Data Schema & Taxonomy

File Path Purpose Persistence
~/.openclaw/config/veille/config.json Stores RSS sources and API keys Permanent
~/.openclaw/data/veille/seen_urls.json Tracks processed URLs 14-day TTL
~/.openclaw/data/veille/topic_seen.json Stores similarity hashes 5-day TTL

Veille Advanced Features

  • Integrated LLM scoring for automated article prioritization and ranking.
  • Modular pipeline architecture compatible with standard CLI tools and Openclaw Skills agents.
  • Security-first design featuring path whitelisting and content sanitization for file outputs.
  • Native support for multiple output channels including Telegram, SMTP, and Nextcloud.

SKILL.md


Loading

Related Openclaw Skills

Featured*