synthesis-context-lifecycle
Installation
SKILL.md
Context Lifecycle Management
The Problem
AI collaborators start every session with zero context. Their effectiveness depends entirely on the quality of the context they receive. For short-lived projects (2-3 sessions), a single context file works. For long-running projects spanning weeks or months, that file grows unboundedly — combining four types of information with fundamentally different lifecycles:
| Information type | Access pattern | Growth pattern | Ideal treatment |
|---|---|---|---|
| Working memory (current state, active tasks) | Every session | Constant | Keep lean, refresh often |
| Episodic memory (session logs) | Rarely after 1 week | Unbounded append | Archive monthly |
| Semantic memory (stable facts, reference) | Most sessions | Slow, update-in-place | Separate file |
| Completed work records | Almost never | Unbounded append | Delete after archiving |
Combining all four in one file means the file grows linearly with session count, with no mechanism for information to leave. This is the classic hot/warm/cold data problem from database engineering, manifesting in AI context management.