continual-learning

原文涵盖的主题

• Continual Learning for AI Coding Agents
• The Loop
• Quick Start
• Two-Tier Memory
• How Learnings Get Stored
• Automatic (via hooks)
• Agent-native (via storememory / SQL)
• Manual (memory files)

Your agent forgets everything between sessions. Continual learning fixes that. Install the hook (one step): Auto-initializes on first session. No config needed. Global (/.copilot/learnings.db) — follows you across all projects: Tool patterns (which tools fail, which work) Cross-project conventions General coding preferences Local (.copilot-memory/learnings.db) — stays with this repo: Project-specific conventions Common mistakes for this codebase Team preferences The hook observes tool outcomes and detects failure patterns: The agent can write learnings directly: Categories: pattern, mistake, preference, toolinsight For human-readable, version-controlled knowledge: Learnings decay over time: Entries older than 60 days with low hit count are pruned High-value learnings (frequently referenced) persist indefinitely Tool logs are pruned after 7 days This prevents unbounded growth while preserving what matters. 1. One step to install — if it takes more than cp -r, it won't get adopted 2. Scope correctly — global for tool patterns, local for project conventions 3. Be specific — "Use semanticconfigurationname=" beats "use the right parameter" 4. Let it compound — small improvements per ses

这里只展示节选，请以完整原文和当前要求为准。