An automated recommendation engine that builds long-term movie and anime taste profiles through iterative feedback loops.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install screen-recommendation-loop
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 screen-recommendation-loop 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 Screen Recommendation Loop?
The Screen Recommendation Loop is a sophisticated framework designed for Openclaw Skills to manage continuous media discovery with minimal user friction. It balances consistency and exploration by recommending one title at a time from diverse sources like Douban and Bangumi. By analyzing feedback from watched, unfinished, or dropped content, this Openclaw Skills agent builds an evolving taste profile to refine future suggestions.
This skill is built for developers and enthusiasts who want a low-maintenance recommendation system that respects their time. It intelligently manages follow-ups and preference weights, ensuring that the AI agent becomes more accurate with every interaction without requiring complex manual configuration.
Screen Recommendation Loop Use Cases
- Automating weekly movie discovery based on historical viewing habits.
- Managing long-term anime watchlists with intelligent, type-based follow-up reminders.
- Refining AI taste profiling using granular feedback signals like dropped or partial views.
- Creating a self-improving recommendation system using curated top lists and constrained randomness.
How Screen Recommendation Loop Works
- The skill picks a candidate title from a mixed pool of sources while applying hard filters for duplicates and rejected tropes.
- It sends a concise recommendation message to the user within the Openclaw Skills environment.
- A follow-up event is automatically scheduled based on the content type: 7 days for movies and 30 days for anime series.
- Upon user feedback, the skill collects the status and optional mood markers using a fixed schema.
- Preference heuristics are applied to update weights, reinforcing positive signals or penalizing disliked narrative patterns.
- The loop continues by picking the next title, incorporating a 15-25% exploration quota to prevent recommendation tunnel vision.
Screen Recommendation Loop Setup
To integrate this into your environment, ensure you have the core framework installed and then add the skill configuration.
# Install the skill via the Openclaw Skills manager
claw skill add screen-recommendation-loop
# Initialize the local state storage (JSON or SQLite)
# Configure your source lists such as Douban Top 250 or Bangumi lists
Screen Recommendation Loop Data Schema & Taxonomy
The Screen Recommendation Loop maintains a clean, structured record of every interaction to ensure persistent learning. The schema is designed to be lightweight and compatible with standard Openclaw Skills data storage.
| Field | Description |
|---|---|
| id | Unique identifier for the recommendation |
| type | content category: movie or anime |
| source | Origin of the recommendation (e.g., Douban) |
| recommendedAt | Timestamp of the initial suggestion |
| followupAt | Target timestamp for the next check-in |
| status | User outcome: watched, partial, not_started, dropped, or rejected |
Screen Recommendation Loop Advanced Features
- Content-type-aware follow-up windows that automate scheduling based on content length.
- Proactive interrupt handling that skips wait times if a user returns feedback early.
- Constrained random ranking that balances preference exploitation with discovery exploration.
- Dynamic preference decay to ensure the model adapts to changing tastes over time rather than overfitting to old data.
SKILL.md
Loading
Related Openclaw Skills
Daily Commits Summary
A specialized tool that analyzes git history to generate human-readable daily summaries grouped by feature and functional area.
Orbcafe Graph Detail AI
A specialized technical skill for building ORBCAFE graph analytics dialogs, searchable detail pages, and configurable AI agent settings.
SaaS Scaffolder
A powerful full-stack generator that builds production-ready SaaS applications with integrated authentication, database schemas, and Stripe billing.
Workspace Health
A specialized utility to identify and resolve nested workspace structures and path configuration errors within the Openclaw environment.
Daum Toy Search
A high-performance search skill for Korean web results featuring Kakao index integration and intelligent reranking for AI agents.
polaris-news
Verified news intelligence providing real-time, bias-scored briefs across 18 specialized verticals for AI agents.
Featured*
