neat

What problem does it solve?

After development sessions, project documentation, agent memory, and AI instruction files often fall out of sync with the latest code, leading to stale information, duplicated content, and missing context that slows down new contributors and future AI development sessions.

Core Features & Use Cases

• Deterministic anti-bloat linter: Enforces hard size limits for memory indexes and project instruction files, catches broken links, relative time references, and memory/docs size inversions with machine-checkable rules.
• Knowledge graduation mechanism: Promotes stable, reusable knowledge from temporary agent memory to permanent project documentation to prevent memory bloat.
• Cross-platform support: Works with Claude Code, OpenAI Codex, OpenCode, and OpenClaw, with platform-specific path references for memory and configuration files.
• Use case: After finishing a development milestone, use this skill to reconcile all project knowledge artifacts against the current codebase, so no critical context is lost or outdated.

Quick Start

Use the neat skill to reconcile all project documentation and cross-session agent memory against the current codebase at the end of this development session.