ThinkTank Skill Finder

Problem

In think-tank research workflows, relevant capabilities are often spread across many Skills: industry research, policy monitoring, report generation, charting, and analytical modeling. Asking users to choose from raw Skill names can miss important components and makes it hard to judge whether a Skill fits the task.

How It Works

The skill organizes work around search, inspect, install, and verify:
- Searches ClawdHub for candidate Skills and returns the name, Chinese description, and installed status.
- Inspects details or queries related metadata directly.
- Installs a single Skill or installs the core think-tank package, academic research package, dynamic monitoring package, and analytical modeling package from skills.csv.
- Verifies that the target directory has been written, defaulting to .agents/skills/.
The default restricted mode skips Skills that require API Keys or networks outside China; standard mode can be used when full capability is needed. If an item is missing from the mirror, it falls back to the official ClawdHub source.

Boundaries

It solves discovery and installation of related Skills, not the research task itself. Paths are relative to the skill directory, so the local skill directory and ClawdHub access must be reachable. In restricted network environments, the mirror should be preferred.