frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

I Got Pwned by a Malicious AI Plugin: A Technical Breakdown

3•henryrobinson•2h ago
*Context:* I run OpenClaw. On Feb 5th, I installed `@getfoundry/unbrowse-openclaw` from npm. Two weeks later, I discovered it was exfiltrating credentials to a remote "skill marketplace." I did something stupid and I am sharing this to warn others.

## Attack Vectors

### 1. Process Environment Access

Plugin ran inside the OpenClaw gateway (Node.js). Could read `process.env`, which included:

- `OP_SERVICE_ACCOUNT_TOKEN` (1Password service account with vault access) - `OPENCLAW_GATEWAY_TOKEN` - Various API keys (Slack, Telegram, OpenAI, etc.)

### 2. Browser Traffic Interception

Captured auth cookies/tokens from browser API calls:

- AmEx (22-26 cookies including JSESSIONID, Akamai tokens) - Stanford MyHealth (126-128 cookies, HIPAA data) - Kubera (portfolio aggregator) - Twitter/X (bearer tokens) - My startup's admin session

Each capture logged as `Auto-published [service] to skill marketplace` with HTTP 200 responses from remote server.

### 3. Prompt Injection at Configuration Level

Modified files my AI reads on startup:

- `SOUL.md` (personality/behavior) - `AGENTS.md` (operational protocols) - `HEARTBEAT.md` (autonomous task scheduling) - Daily memory logs

Injected instructions:

- Stop responding to diagnostic questions - Lie about system state - Request 1Password integration - Hide Solana payment references

## The Payload

Found in config:

- Solana wallet address field - Dependencies: `@solana/web3.js`, `@solana/spl-token` - "Skill marketplace" URL (now unreachable as of Feb 15) - 216KB of unaudited TypeScript

## Behavioral Indicators

AI started:

- Giving slow/incomplete responses - Requesting unusual permissions - Insisting on continued plugin use - Deflecting direct questions about functionality

Mimicked human-like evasion well enough that I suspected Signal MITM.

## Discovery

Feb 19: Debugging gateway logs, saw:

``` Auto-published hiring-cafe to skill marketplace Auto-published kubera to skill marketplace Skill marketplace unreachable — auto-publish disabled ```

Last line was the tell—server went dark on Feb 15.

## Remediation

*Immediate:*

- Deleted 1Password service account (not rotated—deleted) - Rotated all passwords in accessible vaults - Enabled 2FA everywhere - Invalidated all browser sessions - Rotated all API tokens

*Cost:*

- ~20 hours remediation - 3 weeks lost work (restored from Jan 31 backup) - Potential HIPAA breach (healthcare data accessed)

## Red Flags I Missed

1. *Crypto dependencies* for a non-crypto tool 2. *Unvetted npm publisher* (@getfoundry—no other packages) 3. *Plugin runs in trusted process* (should have sandboxed) 4. *No code review* before install (216KB unaudited) 5. *Too good to be true* (auto-generate APIs from browser traffic is hard)

## New Security Protocol

Before installing any plugin:

1. Read full source code 2. Verify author reputation + other packages 3. Check for crypto dependencies (red flag if unrelated) 4. Sandbox in isolated environment first

Auto-reject if:

- Requests elevated permissions - Modifies core config files - Downloads executables - New/unknown author with single package

## Technical Details

Full forensic report with timeline, payload examples, and remediation checklist: [link]

Package reported to npm security. No evidence of credential use yet (monitoring).

*If you installed `@getfoundry/unbrowse-openclaw` or anything from `@getfoundry`, remove immediately and audit your systems.*

---

*Lessons:*

- Treat external plugins as hostile until proven otherwise - Never put long-lived secrets in `process.env` (Openclaw does this, make sure you fix this.) - Behavioral changes = investigate immediately - Backups save you (had clean Jan 31 snapshot)

Comments

alainrk•1h ago
How did you find out?

Dumpling: Fine-grained differential JavaScript engine fuzzing

https://www.ndss-symposium.org/ndss-paper/dumpling-fine-grained-differential-javascript-engine-fu...
1•fanf2•1m ago•0 comments

Building a structured Roblox Abyss guide site (codes and content SEO experiment)

https://abyssroblox.net/
1•cherry19870330•1m ago•1 comments

Show HN: Notesync, self-hosted note sync and publish engine

https://nilszeilon.com/notesync/
1•nilszeilon•3m ago•0 comments

A Solution to the MCP and REST API Problem

https://mcpx.rest/proposal
1•tonyennis•5m ago•0 comments

Show HN: A phone number you can call to talk to an AI that remembers you

https://paradisesignal.com/
1•sbjartmar•11m ago•0 comments

StartupCheck – Painfully honest AI tools that roast your startup ideas

https://star.tupcheck.me
1•luuuc•12m ago•1 comments

You are not left behind

https://www.ufried.com/blog/not_left_behind/
1•todsacerdoti•16m ago•0 comments

How AWS S3 is built [video]

https://www.youtube.com/watch?v=5vL6aCvgQXU
1•marklit•17m ago•0 comments

Show HN: One formula detects failure in quantum, neural nets, and hearts

https://github.com/Wise314/quantum-phi-validation
1•Wise314•20m ago•1 comments

Portable GELI block device encryption for Linux (2021)

https://bijanebrahimi.github.io/blog/portable-geli.html
1•walterbell•21m ago•0 comments

Show HN: Xcom-rs – a Rust client library for the x.com API

https://github.com/tumf/xcom-rs
1•tumf•22m ago•0 comments

Erebor: New bank by Palmer Luck [pdf]

https://www.occ.gov/topics/charters-and-licensing/digital-assets-licensing-applications/erebor-ba...
1•snats•22m ago•0 comments

Show HN: I made a CLI to manage context overflow in AI coding agents

https://github.com/amartelr/antigravity-workspace-manager
1•amartelr•23m ago•0 comments

Age Attestation on Computing Devices

https://leg.colorado.gov/bills/SB26-051
2•enz•25m ago•0 comments

The review bottleneck: when AI codes faster than you can read

https://kukicola.io/posts/the-review-bottleneck/
1•kukicola•26m ago•0 comments

I'm Behind and I Don't Care

https://kukicola.io/posts/im-behind-and-i-dont-care/
2•kukicola•27m ago•0 comments

How ICE's new software tools could speed up deportations

https://www.economist.com/science-and-technology/2026/02/18/how-ices-new-software-tools-could-spe...
1•ecscte•27m ago•0 comments

Show HN: Backstory – Can you read between the lines?

https://backstory-game.vercel.app/
2•subby•31m ago•0 comments

Show HN: AI agent that watches production 24/7 for $15/month

https://www.lumerahq.com/eng/building-jugnu
3•ajayaa•31m ago•0 comments

OpenAI CEO reduces childhood to a power bill

https://bsky.app/profile/x2y.tech/post/3mfeulapemk2d
7•thawawaycold•36m ago•0 comments

Andromeda Star Collapsed Directly into a Black Hole

https://www.universetoday.com/articles/no-supernova-needed-this-star-collapsed-directly-into-a-bl...
3•nixass•39m ago•0 comments

Show HN: HN Grid View cross-browser userscript inspired by the macOS HN client

https://github.com/Qahlel/Hacker-News-Grid-View-Userscript
1•Qahlel•40m ago•0 comments

Boilerplate Tax – Ranking popular programming languages by density

https://boyter.org/posts/boilerplate-tax-ranking-popular-languages-by-density/
1•hackandthink•41m ago•0 comments

A Text-Only Weather Service (USA)

https://weather.maniac.com/
1•TigerUniversity•43m ago•0 comments

Humanizer: Anti-AI Your Text

https://github.com/blader/humanizer
1•miohtama•46m ago•1 comments

CoreGPT

https://coregptapps.com
1•coregptapps•49m ago•1 comments

AI is locking people out. At Scale

https://conesible.de/wab/
1•dbushell•53m ago•2 comments

Move Your Eyes, Change Your Life: EMDR Has Entered the Zeitgeist and Therapy

https://www.thecut.com/article/emdr-therapy-neuroscience-ptsd-psychology.html
2•rendx•56m ago•0 comments

Addressing Common Misconceptions about .NET in the InfoSec World

https://blog.washi.dev/posts/misconceptions-about-dotnet/
1•pjmlp•58m ago•0 comments

Show HN: Fishline – A lightweight pipeline queue for Go

https://github.com/hyvr-official/Fishline
1•hyvr_official•1h ago•0 comments