alexgreensh

alexgreensh / token-optimizer

Public

Audit your Claude Code setup. See exactly where your context window goes and get it back.

agentskills claude-code token-optimization
12
0
100% credibility
Found Mar 03, 2026 at 12 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
HTML
AI Summary

A tool that audits and streamlines Claude Code setups to reduce unnecessary overhead in AI session context usage.

How It Works

1
😩 Feeling crowded chats

Your AI helper conversations start feeling stuffed and slow because of all the extra instructions piling up.

2
🔍 Find the helper tool

You discover this friendly optimizer on GitHub that promises to clean up the clutter and free up space.

3
📥 Easy one-click setup

Run a simple download command and it neatly places itself in your AI helper's home folder.

4
💬 Say the magic words

Open your AI chat session and type a short command to wake up the optimizer.

5
📊 See your savings revealed

Clever mini-helpers quickly scan everything, show a colorful dashboard of waste, and list easy wins to reclaim space.

6
✅ Pick and apply fixes

Review the safe backups and changes, approve what you like, and watch it tidy up automatically.

🚀 Chats fly faster now

Your AI helper has tons more room for real work, responses are sharper, and sessions last longer without glitches.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 12 to 12 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is token-optimizer?

Token-optimizer audits your Claude Code setup to pinpoint context window waste from bloated CLAUDE.md files, unused skills, MCP tools, and rules, reclaiming 5-15% of your 200K tokens per message. Run `/token-optimizer` in a fresh session for a prioritized fix list from six parallel agents, or use the standalone CLI for reports, before/after comparisons, and an interactive HTML dashboard. It backs up everything, shows diffs, and handles progressive disclosure to load heavy content only when needed, like turning workflows into on-demand skills.

Why is it gaining traction?

Unlike manual Claude audits or tools like ccusage that track spend but ignore fixes, this delivers quick wins via agent-driven analysis, usage trends from session logs, and session health checks for stale processes. Developers hook it into SessionEnd for automatic claude audit logs collection into a local DB, spotting idle skills costing 100 tokens each per session. The token optimizer MCP integration and dashboard make optimization visual and repeatable, even headless over HTTP.

Who should use this?

Claude Code power users juggling 50+ skills, multiple MCP servers, and agent teams where overhead multiplies across subagents. Ideal for devs running long sessions or plan mode, tired of early compaction degrading output, or needing github audit log-style tracking for their claude code security audit and token optimizer ai habits.

Verdict

Grab it if Claude Code is your daily driver—install.sh gets you running in seconds with solid docs and zero deps beyond Python 3.8. Low 12 stars and 1.0% credibility score signal early maturity, but the measurement accuracy and backups make it low-risk for reclaiming context fast.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.