Points at your Claude Code session history and tells you where you're wasting money or time. It pulls data from the local analytics API (session logs, token usage, pricing) and looks for patterns like using expensive models on simple tasks, poor cache hit rates, or sessions that dragged on past the point of usefulness. The output is a prioritized table with quantified recommendations: switch model X to model Y here, split this type of task into smaller sessions, these tools keep failing so try alternatives. Honest take: only as good as your session volume. If you've run dozens of sessions, the cost and speed suggestions can be legitimately eye opening. If you're new, it won't have much to work with.
npx -y skills add hoangsonww/claude-code-agent-monitor --skill optimization-suggest --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
sickn33/antigravity-awesome-skills
kubesphere/kubesphere
supercent-io/skills-template