Recommend

What This Skill Does

The Recommend skill is a sophisticated, context-aware engine designed to bridge the gap between user intent and high-quality decision-making. Unlike static recommendation systems, this skill employs a rigorous five-step core loop: Context, Preferences, Research, Match, and Recommendation. It functions by first synthesizing deep user insights from internal knowledge bases, extracting specific preference dimensions (Values, Constraints, History, and Mood), and then cross-referencing these against real-time research data. By prioritizing alignment over simple popularity metrics, the Recommend skill ensures that every suggestion is tailored to the user's current situation rather than generic data trends.

Installation

To integrate this skill into your environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/ivangdavila/recommend

Ensure your OpenClaw environment is updated to the latest version to support the adaptive learning and memory persistence features inherent to this module.

Use Cases

This skill is ideal for scenarios requiring nuanced decision support, such as:

• Professional sourcing: Finding the perfect SaaS tools or vendors based on specific technical constraints.
• Personal planning: Curating dining, travel, or entertainment experiences that align with past feedback and current emotional state.
• Strategic filtering: Navigating large datasets to identify the few high-impact options that meet strict budget or quality requirements.

Example Prompts

1. "I need to find a project management tool for my team of 10. We value extreme ease of use over deep feature sets and are on a strict budget of $5/user/month. What do you recommend?"
2. "Research and recommend three quiet, high-speed coworking spaces in downtown Chicago that are suitable for client meetings and accessible on weekends."
3. "My previous attempts at learning Python were too abstract. Find me a hands-on, project-based curriculum that focuses on data analysis rather than web development."

Tips & Limitations

To get the most out of the Recommend skill, always provide as much context as possible. The system is designed to ask clarifying questions if user signals are insufficient; treating these as a collaborative dialogue will significantly improve accuracy. Be aware of the 'Projecting' trap: the skill is tuned to avoid assuming your tastes mirror common trends. Users should also note that 'Recency Bias' can occur; if your preferences have recently shifted, explicitly state that in your prompt to help the agent recalibrate your user profile. Avoid over-constraining your initial request, as this may lead to zero results; define your must-haves versus your preferences clearly to allow the matching algorithm the space to find the optimal balance.