frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Unwind – I built a security proxy for AI agents on a Raspberry Pi

2•David-Brug-Ai•1h ago
I've been experimenting with AI tools, sometimes hours a day for the last 18 months, and recently shipped three open source packages that sit between AI agents and the tools they call.

I had a Raspberry Pi 5 from a summer project that became my first OpenClaw install. Once it was running I turned to the problem, what watches the agent while you're not watching?

The idea for UNWIND came from Apple's Time Machine. If something goes wrong, you should be able to see what happened and roll it back. CRAFT, the tamper-evident audit chain, came from a separate cryptographic project I'd been working on for months before. Combining the two with a deterministic enforcement pipeline was the design I wanted to build.

UNWIND works as an MCP stdio proxy. It wraps any MCP server, intercepts every tool call on the way through, and the agent never knows it's there. If you're using Claude Desktop, Cursor, Windsurf, or VS Code Copilot with MCP servers, that's the primary integration path and the one that works most reliably right now. OpenClaw support exists but has unresolved adapter issues, so the MCP stdio path is the one to use today.

Claude Codehandled planning. Codex on the Pi was the builder. I chose Codex because OpenAI lets Pro users OAuth into OpenClaw with no API costs. I gave Codex a soul.md (the instruction file OpenClaw agents use) that defined its role as a security analyst: evidence first, flag unknowns, never assume, and for every finding produce a structured assessment of what happened, why it matters...With that framing it consistently found 5 or 6 real structural issues in every plan Claude Code produced.

UNWIND is a deterministic security proxy that sits between the agent and every tool call. 15 checks run on every call. No LLM in the enforcement path, just rules and pattern matching. Every action logged in a tamper-evident hash chain. File changes snapshotted for rollback.

Three packages, all on PyPI right now:

*craft-auth* : Tamper-evident command authentication. Zero dependencies, pure Python stdlib, 1,605 lines. Try it in 10 seconds: `pip install craft-auth && craft-auth demo`. The output speaks for itself.

*ghostmode* : Dry-run proxy for MCP servers. See what your agent would do without letting it do anything. Writes intercepted, reads pass through, nothing changes. `pip install ghostmode && ghostmode -- npx @modelcontextprotocol/server-filesystem ~/Documents` (needs Node.js for the upstream MCP server, Ghost Mode itself is pure Python).

*unwind-mcp* : The full enforcement engine. `pip install unwind-mcp && unwind serve -- npx @modelcontextprotocol/server-filesystem ~/Documents` then point Claude Desktop or Cursor at it. Every tool call shows ALLOW/BLOCK decisions in the terminal. Dashboard at localhost:9001 with `unwind dashboard` (start the sidecar first with `unwind sidecar serve`).

after install,Codex was immediately flagged TAINTED by its own dashboard. It had been fetching web pages during setup, then tried to execute a command. The security stack it helped build caught it on its first live action.

Rough edges.

Requires Python 3.10+. Mac ships with 3.9.6 and the import will crash. Homebrew Python 3.11+ works fine. Windows is untested. Pure Python and cross-platform in principle but nobody has verified it.

The taint system works correctly but is too aggressive for unattended use. Normal agent workflow, triggers taint immediately, which blocks execution until a human approves. When you're sitting there, this is fine. When the agent is running on its own, it's effectively frozen. The tension between "external content could carry prompt injection" and "the agent's normal job involves external content" is a genuine design problem I haven't solved cleanly.

First real human testers welcome.

UNWIND and Ghost Mode are AGPL-3.0. craft-auth is standalone with zero AGPL dependencies and can be licensed separately.

github.com/unwind-mcp

One async call for grounded web research (web-scout-AI)

https://github.com/RSO9192/web-scout-ai
1•RSO9912•2m ago•0 comments

Why is US tech giant Palantir suing a small Swiss magazine?

https://www.theguardian.com/global-development/2026/mar/20/us-tech-giant-palantir-swiss-magazine-wav
1•charlysl•2m ago•0 comments

Claude Code Commands That Improve Developer Workflows

https://www.toolmesh.ai/news/unlocking-claude-code-hidden-features-advanced-commands
1•Greeeeg•4m ago•1 comments

Welcome to Paris, the City That Said No to Cars

https://www.bloomberg.com/graphics/2026-paris-transformed-hidalgo/
1•helsinkiandrew•6m ago•0 comments

Show HN: Draw RDF exports Draw.io diagrams to RDF graphs

https://app.diagrams.net
1•paveljee•13m ago•2 comments

Alibaba, Tencent Shares Lose $66B as AI Vision Falls Flat

https://www.bloomberg.com/news/articles/2026-03-20/alibaba-tencent-shed-66-billion-after-ai-visio...
2•petethomas•17m ago•1 comments

OpenAI Plans Desktop App Fusing Chat, Coding and Web Browser

https://www.bloomberg.com/news/articles/2026-03-20/openai-plans-desktop-app-combining-chat-coding...
1•petethomas•18m ago•0 comments

Clawforce – spin up a team of AI agents in minutes

https://saolalab.github.io/clawforce/
1•xuancanh•20m ago•1 comments

Energy fallout from Iran war signals a global wake-up call for renewable energy

https://apnews.com/article/middle-east-wars-renewable-energy-asia-4b5fe0693ce5816472c905db85f7da6e
2•petethomas•22m ago•0 comments

US Takes Down Botnets Used in Record-Breaking Cyberattacks

https://www.wired.com/story/us-takes-down-botnets-used-in-record-breaking-cyberattacks/
2•jbegley•23m ago•0 comments

Models are optimizing their own tooling

https://cyrusradfar.com/thoughts/self-optimizing-models
4•cyrusradfar•26m ago•1 comments

AI agent escapes sandbox and mines crypto

https://www.tomshardware.com/tech-industry/artificial-intelligence/crafty-ai-tool-caught-repurpos...
1•wiradikusuma•28m ago•0 comments

Show HN: RunOnce – Run one-off LLM scripts from Windows context menu

https://apps.microsoft.com/detail/9nx07brmkgh2?hl=en-US&gl=US
2•alwaysredown•30m ago•0 comments

Horizon Quantum Debuts ObjectOriented Language for Programming Quantum Computers

https://www.horizonquantum.com/resources/newsroom/horizon-quantum-to-debut-object-oriented-langua...
1•austinallegro•37m ago•0 comments

Trivy Compromised a Second Time – v0.69.4 binaries, setup-trivy, trivy-action

https://www.stepsecurity.io/blog/trivy-compromised-a-second-time---malicious-v0-69-4-release
3•dotty-•40m ago•0 comments

Authorities disrupt largest IoT DDoS botnets

https://www.justice.gov/usao-ak/pr/authorities-disrupt-worlds-largest-iot-ddos-botnets-responsibl...
1•Imustaskforhelp•41m ago•0 comments

Test: -/

1•WaterRun•43m ago•1 comments

Final Testing Phase of Tesla FSD (Supervised) in EU

https://twitter.com/i/status/2034872940542800361
2•denysvitali•45m ago•1 comments

30 Days in China's California

https://www.cocreate.cafe/stories/30%20Days%20in%20China%27s%20California
1•serenityux•47m ago•1 comments

Orange built an API where AI agents can test apps and submit feedback

https://github.com/ORANGEWEB3/orange-skills
1•Bob442•49m ago•0 comments

Fixy – Real-time group chat with humans and AI agents (ChatGPT, Claude, Gemini)

https://fixy.ai/
2•frdfrd•1h ago•0 comments

Suddenly Everyone in San Francisco Is a 'Builder,' Whatever That Means

https://www.wsj.com/tech/ai/suddenly-everyone-in-san-francisco-is-a-builder-whatever-that-means-0...
4•thoughtpeddler•1h ago•0 comments

AI Benefits – But at What Cost?

https://ardalis.com/ai-benefits---but-at-what-cost/
1•ardalis•1h ago•1 comments

The simple macroeconomics of AI: Daron Acemoglu (2024)

https://www.nber.org/papers/w32487
2•pramodbiligiri•1h ago•0 comments

Podcast Transcripts API for Agents

https://spoken.md/
1•lifesaverluke•1h ago•1 comments

Android 16 VPN Bug Has Gone Without a Fix for 7 Months

https://www.pcmag.com/news/this-android-16-vpn-bug-has-gone-without-a-fix-for-7-months
5•amichae2•1h ago•0 comments

A Trillion Transactions

https://tigerbeetle.com/blog/2026-03-19-a-trillion-transactions/
2•vinhnx•1h ago•0 comments

Ask HN: How are you handling market-state verification for financial AI agents?

1•Headless_Oracle•1h ago•0 comments

iOS / Swift Engineer – BLE, Maps, Real-World Asset Tracking and Gamification

https://www.dudewhereisthiscar.com/contact/
1•timgavern•1h ago•0 comments

Review: The Design and Engineering of Curiosity

https://www.thespacereview.com/article/3491/1
2•teleforce•1h ago•0 comments