type skillowner Thomas Cherickalcurator Thomas Cherickalrole llm-engineercategory ai-ml/llm-systemsversion 0.1.0status seedupdated 2026-11-15
context-budgeting
Steps
- Measure tokens per section: system, tools, documents, history.
- Rank sections by contribution to output quality.
- Apply trimming, summarisation or retrieval to the lowest-value sections.
- Place stable content first to maximise prompt-cache hits.
Done When
- [ ] Budget table before and after
- [ ] Quality unchanged on the golden set
Registry
| id |
role |
version |
status |
owner |
last_tested |
| SK-LLM-ENGINEER-context-budgeting |
llm-engineer |
0.1.0 |
seed |
Thomas Cherickal |
- |
Links
