memory-management

To decode enterprise chat acronyms and nicknames into actionable context, this Skill resolves internal shorthand, project names, and internal language so Claude understands requests like a colleague would, preserving context across conversations.

The tiered lookup flow resolves shorthand by checking the CLAUDE.md hot cache first, then querying a memory glossary, and finally retrieving deeper context from per-entity profiles stored in memory/projects and memory/people.

Yes, you can update glossaries and personal or project entries for ongoing learning, ensuring that newly introduced acronyms and nicknames are saved to memory for accurate future context retrieval and decoding.

The best way to preserve context across enterprise chat conversations is to translate shorthand requests into full, actionable instructions using a memory glossary and per-entity profiles for people and projects.

A CLAUDE.md hot cache is used for quick decoding of the most-used terms and people, serving as the first tier in the lookup flow before falling back to the broader memory glossary and entity profiles.

Shorthand decoding for task execution may be unnecessary for conversations strictly using standard terminology without internal acronyms, project names, or nicknames that require contextual resolution to a memory glossary.