memory-management

Decode workplace shorthand by maintaining a tiered memory system that resolves internal language into actionable context. It uses a hot cache in CLAUDE.md for quick lookups and a full glossary in the memory directory for comprehensive term and name resolution.

Managing context memory involves keeping CLAUDE.md slim with the top 30 people and terms, while allowing the memory/ directory to grow. A tiered lookup mechanism resolves names by checking the hot cache first, then the full glossary, and finally detailed project directories.

Bootstrap the two-tier memory by loading CLAUDE.md as the hot cache for immediate lookups. Populate the memory/glossary.md file and memory/people, memory/projects, and memory/context directories as new terms, names, and projects appear so Claude can decode requests in real time.

Context decoding remains effective as the glossary grows because the system uses tiered lookups. CLAUDE.md stays slim with the top 30 terms while the memory/ directories expand to reflect new terms and people, ensuring full context resolution without overloading the hot cache.

Claude fails to understand internal shorthand when workplace terminology is not mapped to actionable context. This Skill solves the problem by using a tiered lookup across a hot cache and full glossary to decode acronyms and internal language so Claude responds like a colleague.

Yes, you can use memory/projects/, memory/people/, and memory/context/ directories to store rich, contextual details. The tiered lookup mechanism checks these directories after the hot cache and glossary to fully resolve terms and names for Claude.