gap-finder

Installation
SKILL.md

Gap Finder

Read a paper. Find where the theory is thin.

The goal is not to "critique" the paper in a reviewer sense. The goal is to find publishable theoretical gaps — places where a mathematician could contribute something new. A gap is valuable when: (1) the method works empirically but lacks theoretical justification, (2) the theory exists but under assumptions that don't match practice, or (3) a natural generalization is left open.

Mode selection

  • Full gap scan: run Steps 1–5 below.
  • Weekly lemma autopsy: run the compact protocol below on exactly one lemma.

Weekly lemma autopsy

Use one in-scope lemma from spectral graph theory, community detection, GNN theory, or the user's current area. The habit is one teardown per week, not an exhaustive paper review.

Installs
45
GitHub Stars
2
First Seen
Jun 15, 2026
gap-finder — jurgendn/agent-skills