gremlin-forge

Pattern discovery through skill collision systematically combines disparate reasoning frameworks to create emergent insights. Gremlin-forge applies collision engine logic to your .claude/skills ecosystem, extracting effective strategies and storing them as learnings in .claude/brain for retention and reuse.

Random collisions explore the full design space of possible pattern combinations across all skills. Targeted collisions focus the collision engine on specific skill pairs or domains, giving you control over which reasoning frameworks intersect to generate specialized analysis tools.

Pattern extraction reads SKILL.md frontmatter from your skill ecosystem, identifies codified reasoning strategies, and systematically documents the effective approaches discovered. Results store as Git objects in .claude/brain, creating a searchable record of emergent patterns for future reasoning tasks.

Yes. Collision engine forces conceptual collisions between disparate domains to discover specialized analysis frameworks. For example, you can collide quantum metrics reasoning with consciousness analysis patterns to generate novel frameworks; the Skill formalizes this discovery process systematically.

Discovered patterns are codified, documented via pattern extraction, and stored as Git objects in .claude/brain. This learning retention enables you to recall, refine, and reapply effective reasoning strategies across future projects without repeating discovery work.

Yes. Skill collision requires an existing .claude/skills ecosystem to collide patterns from. The Skill generates new reasoning frameworks by combining patterns already present in your skill library, so a baseline of documented skills provides the raw material for discovery.