continuous-learning

Extract reusable patterns from Claude Code sessions via a Stop hook.

Updated Mar 18, 2026
One-click install
npx skills add https://github.com/freedom909/real-estate-saas --skill continuous-learning-freedom909
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: continuous-learning
Source: https://github.com/freedom909/real-estate-saas/tree/main/.trae/skills/continuous-learning
Command: npx skills add https://github.com/freedom909/real-estate-saas --skill continuous-learning-freedom909

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires jq.

What problem does it solve? Valuable problem-solving patterns discovered during Claude Code sessions are lost when the session ends, forcing you to rediscover the same fixes and workarounds repeatedly. ## Core Features & Use Cases - Automatic Session Evaluation: Runs as a Stop hook at session end to check whether the transcript contains enough messages (default 10+) worth analyzing. - Configurable Pattern Detection: Detects error resolutions, user corrections, workarounds, debugging techniques, and project-specific conventions while ignoring typos and one-time fixes via config.json. - Learned Skill Persistence: Saves extracted patterns as reusable skills in ~/.claude/skills/learned/ for future sessions. - Use Case: After a long debugging session where you resolved a tricky framework quirk, the Stop hook flags the session for pattern extraction so the workaround becomes a permanent learned skill instead of being forgotten. ## Quick Start Add the evaluate-session.sh script as a Stop hook in your Claude Code settings.json and let it evaluate each session for extractable patterns.

Frequently Asked Questions about continuous-learning

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I set up automatic pattern extraction from Claude Code sessions?▼

Add a Stop hook to ~/.claude/settings.json that runs the evaluate-session.sh script with matcher "*". The hook executes once at session end, checks the transcript length, and signals whether patterns should be extracted.

What is the difference between continuous-learning v1 and v2?▼

v1 uses a Stop hook for end-of-session evaluation and extracts full skills directly. v2 uses PreToolUse/PostToolUse hooks for continuous observation, atomic instincts with confidence scoring, and an evolution path from instincts to skills, making it the preferred option for new installs.

Why does the session evaluator skip my Claude Code session?▼

Sessions are skipped when the transcript contains fewer user messages than min_session_length, which defaults to 10 in config.json. Lower this threshold in config.json if you want shorter sessions evaluated.

Does evaluate-session.sh require jq to run?▼

jq is optional. Without it, the script prints a warning to stderr and falls back to built-in defaults: a minimum session length of 10 messages and the learned skills path ~/.claude/skills/learned/.

What pattern types can be detected and ignored?▼

Detected patterns include error_resolution, user_corrections, workarounds, debugging_techniques, and project_specific conventions. Ignored patterns include simple_typos, one_time_fixes, and external_api_issues, all configurable in config.json.