Enterprise data analysis
[COLM'26] SkillLearnBench is the first benchmark for evaluating continual learning methods that automatically generate agent skills.
npx -y skills add cxcscmu/SkillLearnBench --skill enterprise-data-analysisAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
SKILL.md
1.5 KB, as published. Nobody here has run it
name: enterprise-data-analysis description: Analyzing relationships between enterprise data entities (employees, products, reports, teams). Use this skill when a question requires connecting different types of information.
Enterprise Data Analysis
This skill provides a framework for understanding and navigating complex relationships in enterprise datasets.
Relational Mapping
- Product to Team: Look for a "team" or "team_members" field in the product JSON. If it's just names, you'll need to look up their IDs in the
employee.jsonfile. - Product to Competitor: Products usually have a
competitorsarray. This array may contain objects with details like names, strengths, weaknesses, and URLs. - Report to Author/Reviewer: Reports often have
authorsandreviewersfields. These are usually lists of names or IDs. - Insights to Employee: Insights about competitors or products may be linked to specific team members who provided them. Check for "provided_by", "member", or "author" fields within feedback or insight arrays.
Linking Data
- Identify the Source: Determine which product file contains the core information (e.g.,
ContentForce.json). - Extract Primary Keys: Get names or IDs of the relevant people/competitors from the product file.
- Cross-Reference: If only names are found, look up the
employee.jsonto find corresponding IDs. - Consolidate: Bring the findings together to form a complete answer (e.g., a list of IDs).