continuous-learning
Installation
SKILL.md
Default output: return only the result, blockers, and required evidence. Omit preambles, process narration, repeated context, confidence scores, and follow-up offers. Use at most five bullets unless a required artifact or schema needs more.
Continuous Learning
Overview
Learn from sessions in two complementary ways:
- Instinct-based learning — extract reusable patterns, score confidence, promote high-confidence instincts into skills.
- Workspace memory — delegate durable preferences and facts to the
agents-memory-updatersubagent (orchestration only in this skill).
When to Use
- After completing a significant feature or fix (instincts)
- When a session revealed a non-obvious solution worth remembering
- Before starting similar work to apply learned instincts
- When transcript deltas may update
AGENTS.md(workspace memory)