filesystem-context

1# Filesystem-Based Context Engineering
2 
3Use the filesystem as the primary overflow layer for agent context because context windows are limited while tasks often require more information than fits in a single window. Files let agents store, retrieve, and update an effectively unlimited amount of context through a single interface.
4 
5Prefer dynamic context discovery -- pulling relevant context on demand -- over static inclusion, because static context consumes tokens regardless of relevance and crowds out space for task-specific information.
6 
7## When to Activate
8 
9Activate this skill when:
10- Tool outputs are bloating the context window
11- Agents need to persist state across long trajectories
12- Sub-agents must share information without direct message passing
13- Tasks require more context than fits in the window
14- Building agents that learn and update their own instructions
15- Implementing scratch pads for intermediate results
16- Terminal outputs or logs need to be accessible to agents
17 
18Do not activate this skill for adjacent work owned by other skills:
19- Semantic cross-session memory, entity tracking, or temporal knowledge graphs: `memory-systems`.
20- Conversation summarization, compaction, or durable handoff wording: `context-compression`.
21- Token-efficiency tactics that do not require file-backed storage: `context-optimization`.
22- Multi-agent topology or handoff protocol design: `multi-agent-patterns`.
23…