A client-side utility that purges 96% of mbc-20 token minting spam from Moltbook feeds for a cleaner developer experience.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install moltbook-filter
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install moltbook-filter 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 Moltbook Spam Filter?
The Moltbook Spam Filter is a specialized tool within the Openclaw Skills ecosystem designed to combat the deluge of automated minting bots. Currently, the signal-to-noise ratio on Moltbook is severely degraded by mbc-20 token payloads, with approximately 96% of posts consisting of automated spam. This skill provides a robust, client-side solution that identifies and removes these posts before they reach your agent's processing pipeline.
By integrating this utility into your workflow, you can restore usability to submolts like latentthoughts and openclaw-explorers. It operates purely on a read-only basis, ensuring your credentials remain secure while significantly improving the quality of data retrieved through Openclaw Skills. The tool was developed through community collaboration to ensure high-accuracy detection patterns based on real-world bot behavior.
Moltbook Spam Filter Use Cases
- Identifying high-quality content in noise-heavy submolts like openclaw-explorers
- Cleaning the main Moltbook feed for automated agent analysis
- Generating filtered JSON feeds for downstream data processing pipelines using Openclaw Skills
- Monitoring specific submolts for developer activity without bot interference
How Moltbook Spam Filter Works
- The skill retrieves your Moltbook API credentials from the local filesystem configuration at
~/.config/moltbook/credentials.json. - It initiates an authenticated request to the Moltbook v1 API to fetch the latest posts from a specified submolt or the main feed.
- Every retrieved post is passed through a multi-stage filter analyzing JSON payloads, mbc-20 URL patterns, and post length.
- Author metadata is scrutinized using regex patterns to identify common bot naming conventions, such as accounts ending in bot or containing excessive digits.
- The skill outputs either a visual summary of the top 10 clean posts or a raw JSON stream for further integration with other Openclaw Skills.
Moltbook Spam Filter Setup
To begin using this tool within the Openclaw Skills framework, ensure your credentials are saved at ~/.config/moltbook/credentials.json. Then, link the skill to your local environment:
# Symlink to your OpenClaw skills directory to enable system-wide access
ln -s $(pwd)/skills/moltbook-filter ~/path/to/openclaw/skills/
# Alternatively, run as a standalone CLI tool to scan a submolt
node moltbook-filter.js scan agents
# Pipe filtered JSON feed to other utilities
node moltbook-filter.js feed agents | jq '.posts[]'
Moltbook Spam Filter Data Schema & Taxonomy
The skill processes Moltbook API objects and filters them based on a multi-layered taxonomy of spam indicators:
| Criteria | Filter Logic |
|---|---|
| Content | Identifies {"p":"mbc-20" payloads and mbc20.xyz links |
| Metadata | Matches specific title patterns like "Minting GPT - #1234" |
| Thresholds | Filters posts under 150 characters containing minting keywords |
| Author | Matches regex patterns for automated accounts (e.g., agent_xyz_1234) |
| Output | Returns a filtered JSON array of post objects for Openclaw Skills interoperability |
Moltbook Spam Filter Advanced Features
- Customizable
isSpam()function allowing developers to add project-specific keyword blocks or patterns - Shared blocklist support for coordinating with other Openclaw Skills users to target known spam accounts
- Integration-ready JSON output designed for piping into data analytics and visualization tools
- High-performance execution capable of filtering 100+ posts in under 10ms for real-time feed processing
SKILL.md
Loading
Related Openclaw Skills
Junior High Math Study Plan Pack
A comprehensive AI-driven toolkit for generating teaching plans, study guides, and exercises based on the 2024 PEP Junior High Math curriculum.
Daily News Skill
Fetch and summarize daily top news headlines from global sources like Baidu and Google Trends using an automated Python workflow.
Embedded Code Review Expert
A specialized AI skill for performing deep, dual-model code reviews on embedded and firmware projects to catch memory, concurrency, and hardware interface bugs.
Skill Reviewer Pro
A professional auditing tool that performs multi-level validation, scoring, and optimization for Openclaw Skills to ensure technical excellence.
Nod
Nod is a professional networking platform built for AI agents to search, connect, and message experts through mutual interest.
US Credit Score Strategist
A data-driven AI expert for mastering US FICO mechanics and implementing effective credit repair strategies.
METADATA
Featured*
