The new breakdown shows context usage for agents and helps teams see where tokens are going. Cursor explicitly points users toward rules, skills, MCPs, and subagents when diagnosing setup issues.
This is a practical quality-of-life update for anyone tuning agentic workflows. Better visibility into context pressure can reduce wasted tokens, improve outputs, and make debugging much faster.
Open the latest Cursor changelog entry and review the context usage panel for your agent runs. Use the data to trim noisy rules, simplify tool stacks, or split work across smaller subagents.
Read Original Post →