Eve Agent Optimisation

This is forensic analysis for agent runs that went sideways. It walks you through execution logs to find wasted tool calls, wrong turns, and blind alleys, then traces them back to fixable root causes. The diagnostic workflow is solid: compare token counts, map the critical path, quantify waste as a percentage of total calls. Most fixes land in the SKILL.md, where you add specific file paths, explicit approaches, and output formats to eliminate guesswork. It also covers harness and model tradeoffs, but correctly refuses to change those without approval since they're cost decisions. Use this when a job worked but felt inefficient, or when you're trying to squeeze a complex agent down to its leanest possible execution.

skills.sh installs: 159
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add incept5/eve-skillpacks --skill eve-agent-optimisation --agent claude-code

Installs into .claude/skills of the current project.

Files

Eve Agent Optimisation

The goal: get the agent to its objective in the fewest tool calls, fewest tokens, shortest time. Find where it wastes effort and eliminate it.

Hard Rule: Recommend, Don't Change

Never change the harness, model, reasoning effort, or permission policy without asking the user first. These are cost and capability decisions that belong to the project owner. Diagnose, explain the tradeoff, and recommend — then wait for approval.