# Context Lifecycle The context lifecycle is the full operational cycle of AI context from creation through retirement. Context is not a static artifact you build once; it's a living system that requires ongoing management across four phases. ## Phases ### 1. Build Create the initial context layer: master prompts, identity notes, rules, skills, agent configurations, memory systems, and knowledge base structure. This is where most people stop, treating context as a one-time setup. ### 2. Maintain Keep context current as the world changes. Code evolves, goals shift, projects start and end, conventions change. Without active maintenance, context drifts toward [[AI Context Rot]]. Maintenance includes updating rules, pruning stale entries, and syncing context with reality. ### 3. Review Periodically audit whether AI's understanding is still accurate and aligned. Review catches problems that daily maintenance misses: subtle contradictions ([[Context Confusion]]), accumulated noise ([[Context Distraction]]), and slow drift ([[Context Drift]]). Review cadence matters; too infrequent and rot accumulates, too frequent and it becomes overhead. ### 4. Evolve Expand capabilities by adding new skills, refining agents, updating rules as needs change, and incorporating lessons learned. Evolution is intentional change (adding new context), while maintenance is corrective change (fixing existing context). ## Why the lifecycle matters Most context management failures happen because people treat context as a build-once problem. They invest heavily in Phase 1, skip Phases 2-3, and do Phase 4 reactively. The result is [[Context Entropy]]; context that starts clean and gradually degrades. [[Personal Context Management (PCM)]], [[Team Context Management (TCM)]], and [[Enterprise Context Management (ECM)]] all require lifecycle management, but at different cadences and scales. ## Connection to Context-as-Code [[Context-as-Code]] enables lifecycle management by making context changes visible, reviewable, and reversible. Without version control and review processes, lifecycle management is informal and error-prone. ## References - ## Related - [[Personal Context Management (PCM)]] - [[Context Engineering]] - [[AI Context Rot]] - [[Context Entropy]] - [[Context Hygiene]] - [[Context Drift]] - [[Context Confusion]] - [[Context Distraction]] - [[Context-as-Code]] - [[Team Context Management (TCM)]] - [[Enterprise Context Management (ECM)]] - [[Context Management Maturity Model]] - [[Agentic Context Engineering]] - [[Knowledge Decay]]