[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-930":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":16,"stars7d":17,"stars30d":18,"stars90d":15,"forks30d":15,"starsTrendScore":19,"compositeScore":20,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":21,"fork":21,"defaultBranch":22,"hasWiki":23,"hasPages":21,"topics":24,"createdAt":9,"pushedAt":9,"updatedAt":25,"readmeContent":26,"aiSummary":27,"trendingCount":15,"starSnapshotCount":15,"syncStatus":28,"lastSyncTime":29,"discoverSource":30},930,"DontFeedTheAI","zeroc00I\u002FDontFeedTheAI","zeroc00I","Transparent anonymization proxy for AI-assisted pentesting. Strips IPs, credentials, hostnames and PII before they reach any LLM (Claude, OpenAI, OpenRouter). Local Ollama + regex detection. Per-engagement vault.",null,"Python",585,68,9,1,0,4,20,47,12,9.52,false,"main",true,[],"2026-06-12 02:00:20","# DontFeedTheAI\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-blue.svg\" alt=\"License\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.11+-blue.svg\" alt=\"Python\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLLM-Ollama-green.svg\" alt=\"Ollama\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fproxy-FastAPI-009688.svg\" alt=\"FastAPI\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fplatform-Windows%20%7C%20macOS%20%7C%20Linux-lightgrey.svg\" alt=\"Platform\">\n\u003C\u002Fp>\n\nA transparent proxy that strips IPs, credentials, hostnames, and PII from every request before it reaches the AI — and restores them on the way back.\n\n```mermaid\nflowchart TD\n    shell[\"🖥️ Your Shell\\nnmap -sV dc01.acmecorp.local\"]\n    proxy[\"🛡️ DontFeedTheAI\\ndc01.acmecorp.local → srv-0042.pentest.local\\n10.20.0.10 → 203.0.113.47\\nAdmin@Acme2024! → [CRED_XK9A2B3C]\"]\n    api[\"☁️ Anthropic API\\nsees only\\nsrv-0042.pentest.local\\n203.0.113.47\"]\n\n    shell -- \"① real data\" --> proxy\n    proxy -- \"② surrogates only\" --> api\n    api -- \"③ response + surrogates\" --> proxy\n    proxy -- \"④ real data restored\" --> shell\n```\n\n| Layer | Detects |\n|---|---|\n| 🧠 **Ollama (local LLM)** | hostnames, org names, credentials in prose |\n| 🔍 **Regex** | IPs, hashes, tokens, API keys |\n\nBoth run on your machine. Nothing sensitive crosses the boundary.\n\n### Cloud anonymization APIs exist\nbut they mean a second bill and a second third party with your data.\n\nYou're already paying Claude for reasoning. \n\nThe AI doesn't need your real data for that — only the structure and meaning of your questions.\n\n---\n\n| Who | How it helps |\n|-----|-------------|\n| **Pentesters** | Run nmap, mimikatz, bloodhound output through Claude without exposing client infrastructure |\n| **Developers & SREs** | Debug with production data or internal configs in regulated environments |\n| **Legal & consulting** | Anonymize client contracts, case files, or proprietary IP in AI-assisted reviews |\n| **Finance & compliance** | Analyze reports or audit scripts without exposing account details |\n| **Researchers** | Query LLMs on confidential datasets |\n\n---\n\n## Why not just Ollama or Claude directly?\n\n**❌ Cloud anonymization API + Claude** — two bills, two third parties. Your sensitive data still leaves the machine, just through more hands.\n\n```mermaid\nflowchart LR\n    s0[\"🖥️ Your Shell\\nreal data\"] --> a0[\"☁️ Anonymization API\\nsees everything\\nbill #1\"]\n    a0 --> c0[\"☁️ Anthropic API\\nbill #2\"]\n```\n\n**❌ Ollama alone** — your data never leaves the machine, but Ollama has no awareness of what's sensitive.\nIt reasons on whatever you paste: real IPs, real credentials, real hostnames.\n\n```mermaid\nflowchart LR\n    s1[\"🖥️ Your Shell\\nreal data\"] --> o1[\"🧠 Ollama\\nno interception\\nreasons on real data\"]\n```\n\n**❌ Claude directly** — best reasoning quality, but everything lands in Anthropic's infrastructure.\nReal client IPs, credentials, org names in their API logs — one policy change or breach away from a problem.\n\n```mermaid\nflowchart LR\n    s2[\"🖥️ Your Shell\\nreal data\"] --> c1[\"☁️ Anthropic API\\nsees everything\\nlogs your real data\"]\n```\n\n**✅ DontFeedTheAI** — Claude's reasoning, Ollama's local detection, nothing sensitive crosses the boundary.\n\n```mermaid\nflowchart LR\n    s3[\"🖥️ Your Shell\\nreal data\"] --> p[\"🛡️ DontFeedTheAI\"]\n    o2[\"🧠 Ollama\\nlocal detector\\nnever leaves machine\"] --> p\n    p --> c2[\"☁️ Anthropic API\\nsees only surrogates\"]\n```\n\n→ See [docs\u002Farchitecture.md](docs\u002Farchitecture.md) for the full technical breakdown.\n\n---\n\n## Quick Start\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fzeroc00I\u002FDontFeedTheAI\ncd DontFeedTheAI\npip install -r requirements.txt\npython3 wizard.py\n```\n\nThe wizard asks everything — engagement name, where to run it, VPS address, model — then deploys, opens the tunnel, and launches Claude with the proxy active. Works on Windows, macOS, and Linux.\n\n```bash\npython3 wizard.py --help   # all available commands\n```\n\n---\n\n## Docs\n\n| Doc | About |\n|--|--|\n| [Architecture](docs\u002Farchitecture.md) | Two-layer pipeline, what gets anonymized and what doesn't, config reference |\n| [Contributing](docs\u002Fcontributing.md) | How to add fixtures, run the improvement loop, open areas |\n| [Threat Model](docs\u002Fthreat-model.md) | What this protects against, what it doesn't, limitations, roadmap |\n\n---\n\n## Verifying coverage & contributing improvements\n\nTwo tools ship with DontFeedTheAI to help you validate coverage and extend it.\n\n**Visual audit** — open in browser while the proxy is running:\n\n```bash\npython3 wizard.py tunnel --audit\n```\n\nShows every `ORIGINAL → SURROGATE` mapping logged during the session, filterable by entity type (DOMAIN, CREDENTIAL, TOKEN, HASH…) with per-request timing breakdown. Use it to spot leaks at a glance instead of grepping logs.\n\n![audit dashboard](docs\u002Faudit-screenshot.png)\n\n> The audit page is a **debug tool**. It exposes the full surrogate → original lookup table, which is why it only runs behind the SSH tunnel. Making this write-only (no reverse lookup over HTTP) is on the roadmap — see [Threat Model](docs\u002Fthreat-model.md).\n\n**Testing the full pipeline** — requires Ollama running:\n\n```bash\npython3 wizard.py test --integration\n```\n\nRuns all 53 fixtures through the complete pipeline (LLM + regex) and asserts zero leaks. Without `--integration`, the LLM is mocked and only the regex layer is validated — useful for fast iteration but not a substitute for the full run.\n\n**Auto-improvement loop** — regex layer only, no Ollama required:\n\n```bash\npython3 wizard.py improve --cycles 3\n```\n\nRuns all fixtures through the regex layer, reports leaks and false positives, and tells you exactly which strings slipped through. The contribution cycle is: add a fixture for a real tool you use → run the loop → add a regex pattern for each leak → repeat. See [Contributing](docs\u002Fcontributing.md).\n\nThe two commands complement each other: `improve` tightens the regex floor fast; `test --integration` confirms the full pipeline holds.\n\n---\n\n## A note from the author\n\n> I'm a pentester, not a software architect.\n>\n> This wasn't built to be innovative — there are already cloud APIs that do LLM-based anonymization. But that means sending your data to yet another third party, and I refuse. If you work in security, you already know why.\n>\n> I built this so the architecture would be available to everyone, and so the community could help expand its effectiveness for free. You're paying for context processing — the AI doesn't need your real data for that.\n>\n> — *zeroc00I*\n\n---\n\n## Star History\n\n[![Star History Chart](https:\u002F\u002Fapi.star-history.com\u002Fsvg?repos=zeroc00I\u002FDontFeedTheAI&type=Date)](https:\u002F\u002Fstar-history.com\u002F#zeroc00I\u002FDontFeedTheAI&Date)\n\n---\n\n## License\n\n[MIT](LICENSE)\n","DontFeedTheAI 是一个反向代理工具，用于在数据到达Anthropic的Claude AI之前匿名化敏感的渗透测试数据（如IP地址、哈希值、凭证、主机名和个人身份信息）。它采用双层检测机制：本地Ollama语言模型和正则表达式安全网，确保敏感信息不会泄露。此外，它还支持每次参与的保险库以及自我改进的反馈循环。该工具适合于需要保护敏感信息的各种场景，包括渗透测试、开发与运维调试、法律咨询、财务合规分析以及基于机密数据集的研究工作。通过使用Python编写，并基于FastAPI构建，DontFeedTheAI可以在Windows、macOS和Linux平台上运行。",2,"2026-06-11 02:40:19","CREATED_QUERY"]