continuous-learning

To capture repeated patterns, this Skill uses hook-based observation to record prompts, tool calls, and outcomes, then converts those repeated behaviors into atomic, confidence-scored instincts for reuse.

Yes, you can keep learned coding instincts isolated. This Skill applies project scoping to keep project-specific behaviors isolated, while promoting broadly useful habits into global scope.

Confidence scoring evaluates repeated corrections, resolutions, and workflows captured during sessions to assign a reliability score. This ensures only verified, repeated behaviors become atomic, confidence-scored instincts.

The best way to export promoted instincts is through the built-in evolution workflows. These workflows support exporting, importing, promoting, and clustering instincts into skills, commands, or agents.

Yes, persistent observation storage is required. The instinct extraction process relies on continuous hook-based observation data, project detection, and exportable instinct records to function across multiple coding sessions.

Learned behaviors may not apply globally because of project scoping. This Skill intentionally isolates project-specific behaviors while only promoting broadly useful habits into global scope to prevent incorrect applications.