Implements a persistent file-based working memory system to maintain context and track progress during complex multi-step tasks.
This skill provides a structured methodology for managing long-running agentic tasks by offloading AI context to persistent Markdown files on disk. By maintaining dedicated files for task plans, research findings, and progress logs directly within your project directory, it prevents context loss during long sessions and ensures goal alignment. It features a robust '3-strike' error protocol and specific rules for saving information, making it ideal for complex architectural changes, research-heavy tasks, and multi-step development workflows where standard context windows might become overloaded.
主要功能
01Automated task planning and phase tracking logic
0246 GitHub stars
03Three-strike error resolution and escalation protocol
04Standardized research logging via findings.md
05Context refresh mechanisms to prevent goal drift
06Persistent markdown-based working memory system
使用场景
01Complex multi-step architectural refactoring
02In-depth technical research and dependency analysis
03Building new features from scratch spanning multiple files
