{
  "markdown": "# AgentSkin Suite\n\n[![npm version](https://img.shields.io/npm/v/agentskin.svg)](https://npmjs.org/package/agentskin)\n[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)\n\n**The complete token optimization stack for AI coding agents.** Three complementary tools that eliminate the \"Token Tax\" across the full agent cycle — API responses, CLI output, and agent replies.\n\n## The Suite\n\n| Component | What It Does | Savings | Creator |\n|-----------|-------------|---------|---------|\n| **AgentSkin SSS** | Semantic JSON pruning via MCP | 60–88% on rich APIs | Shawn Nichols Sr. |\n| **Tokenjuice CLI** | Rule-driven terminal output compaction | Up to 99.97% on large outputs | [Vincent Koc](https://github.com/vincentkoc/tokenjuice) (MIT) |\n| **Caveman** | Output compression via prompt engineering | 65% output reduction | [Julius Brussee](https://github.com/JuliusBrussee/caveman) |\n\n## Quick Start\n\n```bash\n# 1. AgentSkin MCP server\nnpx -y agentskin@latest\n\n# 2. Bash hook (transparent CLI optimization)\nsource .agents/hooks/bash-optimizer.sh\n\n# 3. Caveman — already active via AGENTS.md\n```\n\n## By the Numbers\n\n| Metric | Value |\n|--------|-------|\n| GitHub API savings | **88.3%** (1,544 → 180 tokens) |\n| `ls -laR` directory listing | **99.97%** (3.19M → 897 chars) |\n| Caveman output compression | **65%** average |\n| Caveman memory compression | **46%** smaller |\n| MCP tools | **7** across unified server |\n| Combined test suite | **4,695 tests**, 274 files — 100% passing |\n| Pipeline throughput | **3,030 fixtures/sec** (0.33ms avg) |\n| Net session savings | **17.1%** (zero overhead via bash hook) |\n\n## Tools\n\n### AgentSkin Suite MCP (7 tools, unified)\n`fetch_optimized_data` · `skin_reasoning` · `classify_url` · `strip_ansi` · `reduce` · `estimate_tokens` · `apply_json_semantic`\n\n### Caveman Skills (6 skills)\n`caveman` · `caveman-commit` · `caveman-review` · `caveman-stats` · `caveman-compress` · `cavecrew`\n\n## Documentation\n\n| Doc | Description |\n|-----|-------------|\n| [OVERVIEW.md](OVERVIEW.md) | Project overview and quick start |\n| [ARCHITECTURE.md](ARCHITECTURE.md) | Technical architecture and data flow |\n| [STATUS.md](STATUS.md) | Phase completion and test coverage |\n| [BENCHMARK.md](BENCHMARK.md) | Live API compression benchmarks |\n| [USAGE.md](USAGE.md) | Full usage guide — hook, caveman, MCP |\n| [THREE-LAYER-RULES.md](THREE-LAYER-RULES.md) | Rule config override semantics |\n| [AUTHORS.md](AUTHORS.md) | Credits and component details |\n\n## Security\n\n- **SSRF Protection:** Blocks private IPv4/IPv6 ranges and cloud metadata services\n- **Rate Limiting:** 60 req/min sliding window\n- **Input Validation:** Zod schemas for all tool inputs\n- **Processing Timeout:** 30-second limit\n\n## Credits\n\n| Creator | Contribution |\n|---------|-------------|\n| **Shawn Nichols Sr.** (Nichols Transco LLC) | AgentSkin SSS protocol, MCP server, Suite integration |\n| **Vincent Koc** | Tokenjuice — [MIT License](https://github.com/vincentkoc/tokenjuice) |\n| **Julius Brussee** | Caveman — [github.com/JuliusBrussee/caveman](https://github.com/JuliusBrussee/caveman) |\n\n## Website\n\n[agentskin.dev](https://agentskin.dev) — Protocol specification, examples, FAQ, and whitepaper.\n\n---\n\n*© 2026 Nichols Transco LLC.*\n",
  "bytes": 3282,
  "sha": "f2de3ec4f99450a421b6f218067f523ef057edca6c9699f1b98df9be72c150f8",
  "repo_slug": "shawn5cents/agentskin",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_shawn5cents_agentskin_eabfcd6c/readme"
}