Skillsync

Developers using AI coding agents such as Claude Code, Codex, and Cursor often solve the same problem twice because past sessions where they already figured something out are not surfaced again. Today that past work sits in scattered session logs with no easy way to search across prompts, replies, reasoning, commands and file paths. When an agent hits a familiar bug or task, it starts from scratch instead of building on prior fixes, costing developers time re-explaining context or re-debugging issues already solved.

Skillsync is a macOS app that installs a plugin into existing coding agents and reads the sessions those agents already keep, without requiring any indexing or setup. On every prompt, it checks in under a second whether past work is relevant, and if so searches every session in scope, then hands the agent a short recollection with sources so it can continue rather than restart. A global search, triggered from anywhere with a keyboard shortcut, lets users browse sessions, resume them, or carry a conversation into a different coding agent. It also offers a workspace feature so teams can share specific sessions and a skill library, while everything stays private on the user's Mac by default and is processed with zero data retention unless explicitly shared. A separate CLI (skl) supports Linux, Windows and remote servers.