On-Demand Context
Use of scoped, on-demand skill files rather than always-on document loading.
What it measures
Without skill files, every session either re-explains context manually (expensive) or starts without it (risky). Skills built from successful runs dramatically outperform speculative ones.
Do you have skill files for recurring workflows, built from successful runs rather than written speculatively?
The four diagnostic questions
- 01You use skill files so the model pulls detail only when needed, rather than loading everything at session start.
- 02Skills are built from successful runs — you walk the model through a workflow first, then ask it to write the skill.
- 03Skill files include real failure cases and error-handling patterns, not just the happy path.
- 04You iterate skills after failures: feed the error back, fix the workflow, update the skill before the next run.
What it looks like when it works
Mature skills discipline. The next lever is recursive refinement — skills should get tighter with each iteration.
What goes wrong
You're likely loading everything at session start. Build skill files from successful runs — not templates. Context on demand, not context always-on.
Critique & references
The skill-as-needed pattern aligns with token-optimization research: concise, targeted context components measurably improve task accuracy versus monolithic always-on instructions. Pull, don't push.