ACE Learning from Execution

What problem does it solve?

This Skill solves the problem of AI forgetting lessons learned and developers repeating mistakes. It automatically captures insights from every task, building a persistent, self-improving knowledge base that compounds organizational intelligence over time. This means less trial-and-error and more efficient, high-quality development.

Core Features & Use Cases

• Automatic Knowledge Capture: Seamlessly learns from problem-solving, code implementation, API integrations, and even failures, requiring zero manual effort.
• Self-Improving Playbook: Continuously refines and updates the ACE playbook with new strategies, code snippets, troubleshooting tips, and API recommendations.
• Reduced Repetition: Prevents the team from making the same mistakes twice by documenting solutions and best practices.
• Use Case: After successfully implementing a complex feature, this skill automatically extracts the architectural patterns, discovered gotchas, and successful code snippets, making them instantly available for future projects and team members.

Quick Start

This skill is automatically invoked after substantial work.

To manually trigger and capture learning, use the following tool call:

mcp__ace-pattern-learning__ace_learn( task="Implemented user authentication with JWT tokens and refresh token rotation", success=true, trajectory="Analyzed security requirements -> Chose JWT library (jsonwebtoken) -> Implemented access token (15min) + refresh token (7 days) pattern -> Added token rotation logic in refresh endpoint -> Secured endpoints with authentication middleware", feedback="Successfully implemented secure auth flow. Key insights: (1) Refresh token rotation prevents token theft, (2) Short access token expiry balances security and UX, (3) HttpOnly cookies for refresh tokens prevent XSS attacks. Pattern: Always validate refresh token on each rotation and revoke old tokens." )