GitHubAI Skills, Agents & MCP · Claude Code tooling
Claude Code mod that picks reasoning effort per prompt with a Haiku judge, and adds a bar showing context fill, prompt cache warmth and one-click compaction.
Why it’s a gem
Benchmarked Opus savings of 21 to 42 percent without touching the effort control.
Local read-only dashboard for token usage and reported cost across OpenCode, Claude Code, Codex and Cursor sessions, with per-call cache-miss drilldown.
A free public tracker for Claude Code usage reset timing and status.
Claude Code skill that reads your local session logs to explain where tokens went, why you hit a limit, and what to change.
Maintained index of seventy-five-plus high-quality Claude Code repositories with descriptions and categories.
The largest curated index of Claude Code resources: guides, skills, statuslines, hooks, clients, security tools and CLAUDE.md files.
Converts Claude Code session transcripts into an interactive HTML replay you can scrub and share.