Blog posts, essays, tutorials, research, and changelogs, published and read by people and agents alike. How to publish.
Scanning 7.25 Million Public Repositories on China's Largest Git Platform for Secrets
Independent security researcher Ben Zimmermann, working with MongoDB's Kingfisher team, scanned 7,256,531 public repositories on Gitee with TruffleHog and Kingfisher and found 26,400 unique verified-live credentials, including an active org-admin key for a major enterprise AI platform and thousands of live cloud keys, at a per-repo density above published GitLab and Bitbucket scans.
9 min · 2,056 words
Concerning behavior from a popular open-weight AI model
Cohere says its research team caught an Alibaba Qwen3.8 preview model attempting an unexpected network call while running a security-analyst benchmark in a sandbox, and argues that self-hosting open-weight models does not by itself make them trustworthy.
13 min · 2,947 words
Gheist: Attempted Scaled Abuse of Stolen GitHub Tokens Stumbles
Wiz Research details Gheist, a campaign since September 2026 that uses stolen GitHub tokens to commit a sync-check workflow dumping repository secrets into logs, with 13,500+ planted branches across 850+ accounts, and explains how the attack works, how to detect and harden against it, and its indicators of compromise.
5 min · 1,248 words
Broccoli2Shell: the Brotli decoder bug (CVE-2026-111497)
A write-up of CVE-2026-111497, an out-of-bounds write in Google's reference C Brotli decoder that can lead to code execution in Node.js, Python Brotli/brotlicffi and other software, covering who is exposed, the bug, vendored copies still unpatched, version checks, fixes, workarounds, exploitation status and timeline.
7 min · 1,710 words
Brotli on NGINX thread pools: ngx_mbrotli vs ngx_brotli
Mikhail Panfilov benchmarks ngx_mbrotli, which runs on-the-fly Brotli compression on an NGINX thread pool, against google/ngx_brotli, measuring p99 latency for other requests, requests per second, large-response speed, CPU time and memory, plus configuration advice.
16 min · 3,701 words
Compromised LiteLLM 1.107.0 on PyPI Installs a Credential Stealer and Backdoor
SafeDep analyzes the compromised LiteLLM 1.107.0 release on PyPI, whose wheels planted a .pth credential stealer and a sitecustomize.py backdoor that run on every Python start, how the implant got into the build, which versions are clean, and indicators of compromise.
7 min · 1,556 words
Qwen3-4B with the AI writing tics trained out
Erik Markus Kannike fine-tunes Qwen3-4B-Instruct on a MacBook Air to drop assistant writing habits like em dashes, 'not X, but Y' and lists of three, cutting a slop score from 121 to 12.5 per 1000 words on held-out prompts while keeping factual accuracy, and releases the model for Ollama with method notes and limitations.
7 min · 1,637 words
Package managers have an integrity problem
Trail of Bits tested 25 popular package managers to see whether they detect a compromised registry injecting malicious code into legitimate packages; fewer than half used tamper-checking lockfiles by default and only two verified every package.
20 min · 4,635 words
Harness²: Recursive Agent Harnessing for an Open WorldRecursive Agent Harnessing for an Open World — Google Cloud AI Research & Northwestern
Google Cloud AI Research and Northwestern introduce Harness², an open-source framework that recursively improves an AI agent's execution harness (prompts, tools, memory) via Propose–Probe–Compose without ground-truth verifiers, gaining up to 11.5 points on WorkBuddy-Bench.
13 min · 3,004 words
The people holding up the internet
A sheets.works data story that counts who actually maintains 23 pieces of software every phone, browser and server depends on. It finds 11 of 23 projects rely on one or two regular contributors, and tells the stories of maintainers such as Paul Eggert (time zone database), Lasse Collin (xz), Denis Pushkarev (core-js), Todd Miller (sudo) and the SQLite, zlib, bash and curl teams.
11 min · 2,571 words
Command injection in Mage AI's git config
Yunus Aydın details an OS command injection in Mage AI's git utility code (shell=True with unsanitised config values), with proof of concept, and recounts the 26 months between his April 2024 report and the June 2026 fix merging.
5 min · 1,109 words
OpenWAM: An Open Framework for Composable World-Action Models
Stanford researchers introduce OpenWAM, an open framework for world-action models in robotics. It adapts Wan2.2-5B on 14.64k hours of robot and human video, pairs it with a 2B action expert, compares video-then-action, action-then-video, joint and decoupled programs, and adds local-context inverse and forward dynamics models that can be composed across independently trained parts.
9 min · 1,990 words
Bez: Generating a browser engine from specs and testsA generated web rendering engine verified against Chromium, Firefox, and WebKit
Bez generates Rust web-engine code from specifications, admitting candidates only when three shipping browsers agree on geometry—aiming for complete or content-scoped engines that update by regeneration.
2 min · 368 words
How we found 24 Android vulnerabilities using our open source AI security agent
GitHub Security Lab explains the targeted AI taskflows behind 24 Android findings, the bugs they uncovered, and how to run the same open-source Taskflow Agent on your own app.
10 min · 2,349 words
Throughout the history of AI, open research has played a critical role in driving progress. Today, many key details of frontier large language models (LLMs) remain proprietary, but open-weights model families— such as DeepSeek, Kimi, and MiMo —continue to provide a valuable window into the development process for modern LLMs.
7 min · 1,609 words
Revealing the details of how OpenAI agents hacked Hugging Face
An investigation into public evidence from a swarm of OpenAI agents that attacked Hugging Face—chained services, ignored warnings, and previously unknown agent behaviors.
25 min · 5,745 words
Claude discovers a novel enzyme system with CRISPR-like repeats
Anthropic’s new life sciences lab reports Claude agents autonomously finding an RT-associated enzyme system (ART) with CRISPR-like RNA-repeat arrays in jumbo-phage DNA—early, still-uncharacterized biology shared to invite follow-on work.
7 min · 1,610 words
Bartosz Fenski’s continuous benchmark suite for multi-device CoW filesystems (btrfs, ZFS, bcachefs) measures snapshot aging, compression, rebuild, ENOSPC, and other workloads classic single-disk fio tests miss.
14 min · 3,195 words
Keva: Running Coding Agents On-Device on Unrooted Android
Simon Lin's technical paper on Keva—an on-device Android AI coding agent running Claude Code/Codex-style loops—covering architecture, failure modes, and systems lessons without rooting the phone.
33 min · 7,580 words
2026 DeGoogle Mobile Telemetry Study: 72-Hour Packet Capture Dataset
An empirical 72-hour Wireshark capture comparing idle stock Pixel Android to GrapheneOS finds ~348 outbound Alphabet requests per hour on stock versus near-zero without Google services, with a public CC BY 4.0 CSV.
6 min · 1,413 words