continuous-learning

name: continuous-learning description: Automatically extract reusable patterns from Claude Code sessions and save them as learned skills for future use.

Continuous Learning Skill

Automatically evaluates Claude Code sessions on end to extract reusable patterns that can be saved as learned skills.

How It Works

This skill runs as a Stop hook at the end of each session:

1. Session Evaluation: Checks if session has enough messages (default: 10+)
2. Pattern Detection: Identifies extractable patterns from the session
3. Skill Extraction: Saves useful patterns to ~/.claude/skills/learned/

Configuration

Edit config.json to customize:

{
  "min_session_length": 10,
  "extraction_threshold": "medium",
  "auto_approve": false,
  "learned_skills_path": "~/.claude/skills/learned/",
  "patterns_to_detect": [
    "error_resolution",
    "user_corrections",
    "workarounds",
    "debugging_techniques",
    "project_specific"
  ],
  "ignore_patterns": [
    "simple_typos",
    "one_time_fixes",
    "external_api_issues"
  ]
}

Pattern Types

Hook Setup

Add to your ~/.claude/settings.json:

{
  "hooks": {
    "Stop": [{
      "matcher": "*",
      "hooks": [{
        "type": "command",
        "command": "~/.claude/skills/continuous-learning/evaluate-session.sh"
      }]
    }]
  }
}

Why Stop Hook?

• Lightweight: Runs once at session end
• Non-blocking: Doesn't add latency to every message
• Complete context: Has access to full session transcript

Related

• The Longform Guide - Section on continuous learning
• /learn command - Manual pattern extraction mid-session

Sudocode Integration

Patterns → Spec Feedback

When patterns are extracted that relate to a spec, they can become feedback:

Pattern detected: "Error handling for OAuth requires PKCE flow"
Related spec: [[s-auth]] (Authentication System)

→ Automatically suggest:
  add_feedback to_id=s-auth type=suggestion
  content="Consider adding PKCE requirement to OAuth section.
  Discovered during implementation that this is required for
  security compliance."

Learning from Feedback Patterns

Analyze feedback provided across sessions to identify:

1. Common suggestions → May indicate spec template gaps
2. Repeated challenges → May need skill documentation
3. Design decision patterns → Could become decision specs

Session Evaluation with Sudocode Context

When evaluating sessions, include sudocode activity:

{
  "patterns_to_detect": [
    "error_resolution",
    "user_corrections",
    "workarounds",
    "debugging_techniques",
    "project_specific",
    "spec_clarifications",      // NEW: Ambiguities found in specs
    "feedback_improvements",    // NEW: Better feedback patterns
    "dependency_discoveries"    // NEW: Missing issue dependencies
  ]
}

Extracted Patterns as Specs

Significant patterns can be promoted to formal specs:

Pattern: "Always use PKCE for OAuth mobile flows"
↓
Consider creating:
  upsert_spec title="OAuth Security Requirements"
  description="Patterns learned from implementation..."