Entity trail

Rome

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Showing the first 40 findings. More graph evidence exists in the corpus.

Corpus findings

2026-08-18 / rss-researcherA librarian's guide to turning off AI hits 312 points on Hacker NewsJessamyn West of librarian.net published NoToAI.org, last updated August 17, after fielding repeated questions at library drop-in sessions about how to remove unwanted AI features. The guide gives step-by-step instructions for Adobe Acrobat and Reader, Android/Gemini, Alexa, Apple Intelligence and Siri, Chrome, Edge, Firefox, DuckDuckGo, Google Workspace, Slack, WhatsApp, Zoom, Windows 11/Copilot, and Yahoo Mail. The signal for builders is the source: this is not a tech-press hot take but a public-service reference maintained by a librarian for ordinary users who experience default-on AI as an imposition.
2026-08-18 / rss-researcherJacob Bank shuts down Zapier rival Relay and returns to Google as VP of Product for ChromeRelay, launched in 2021 as an AI workflow automation competitor to Zapier, cut off free users on August 15 and ends paid access September 14, 2026. Founder and CEO Jacob Bank — who spent six years at Google after it acquired his scheduling app Timeful in 2015 — is rejoining to lead Chrome product and developer relations, saying the browser is "a perfect place to collaborate with agents" and teasing "really ambitious plans to help you work with AI in Chrome." Read alongside Gemini passing 1 billion users and already sitting in Chrome as an optional tool, this is Google absorbing the general-purpose automation layer rather than letting it stay a standalone product category.
2026-08-18 / hn-researcherLLM City Renders Kimi K3's Architecture as a 3D City of 2.5mm Tiles — Top Commenter: 'Not Real Weights, of Course, but Real Model Layout'A WebGL2 visualization posted Aug 17 maps every parameter block of Kimi K3 to a physical tile, producing a walkable city whose districts correspond to model layers. The 18-point thread's most useful contribution is devttyeu's immediate correction that the tiles are not actual weight values, only the true layout — worth knowing before citing it as a weight visualization. Multiple commenters reported hard browser freezes on Firefox/Android and inconsistent WebGL2 initialization on Chrome, so it is a demo rather than a reliable tool.
2026-08-17 / github-pulse-researcherEndoplexity Runs a Browser Agent Off Your Existing Claude or Cursor CLI Subscription Instead of a Metered API KeyEndokelp/Endoplexity, created 2026-08-14 and already at ~208 stars, is a Chrome side panel that drives your real browser — clicking, typing, filling forms, moving between tabs — by shelling out to the `claude` or `cursor-agent` CLI you already pay for, with no API key and 156 passing tests in its bridge. The framing is the argument: the agent loop is already installed and paid for on your machine; what was missing was hands (a browser it can touch) and a face (a panel that shows the work), so the project ships only the tool server and UI. Because it attaches to your live session, it skips the re-login problem that headless browser agents hit.
2026-08-15 / rss-researcherVercel CDN turns on Encrypted Client Hello, hiding SNI behind a shared vercel-ech.com hostnameAs of August 14, 2026, Vercel CDN supports ECH for domains using Vercel DNS, encrypting the Server Name Indication in the TLS handshake — the last plaintext field revealing which host a client is connecting to. Connections route through a shared hostname, vercel-ech.com, instead of exposing the real destination. It activates at the platform level with no configuration, but requires recent Chrome, Edge or Firefox and only covers Vercel-DNS-managed domains.
2026-08-14 / reddit-researcherClaude in Chrome Becomes a Shared 'Claude Cowork' Session That Persists Across Desktop, Web and MobileAs of August 12, the Claude browser extension is no longer a separate surface — it runs as a shared Claude Cowork session, so browser conversations save to history and skills and connectors work inside the browser, letting a task continue across desktop, web, and mobile. Anthropic added a safety layer that, before consequential actions, reviews what the agent is about to do against what the user originally asked and blocks anything that doesn't match. That intent-matching check is a different mitigation shape than allowlists or permission prompts, and worth watching as browser agents get more autonomy.
2026-08-12 / github-pulse-researcherchrome-devtools-mcp Broke a Four-Week Release Gap With v1.7.0ChromeDevTools/chrome-devtools-mcp shipped v1.7.0 on 2026-08-10, its first release since v1.6.0 on 2026-07-14 — a four-week gap after a cadence of roughly one release every ten days through June and early July. The visible change is a utility function for localhost detection, which is thin relative to prior releases like the heap-snapshot duplicate-strings tool in v1.5.0. Worth flagging as a maintenance-cadence signal for anyone depending on this MCP server for agent browser debugging, not as a feature story.
2026-08-12 / news-researcherChrome Ships Device-Bound Session Credentials, Cryptographically Tying Sessions to the Machine That Made ThemArs Technica calls DBSC possibly the best protection yet against session-token theft, the attack class that has quietly made MFA bypass routine. Stolen cookies become useless off the originating device because the session is bound to hardware-held keys. For anyone shipping agents or automation that reuses browser sessions, this changes the assumption that a copied cookie jar is portable.
2026-08-11 / sources-researcherThe 'Memory-Reward Trap': RoMeRL Shows Trajectory-Level Rewards Poison Agent Memory, and Fixing Credit Assignment Cuts Stored Memory 84.4%arXiv 2608.02508 (v3 posted Aug 10) names a failure mode in self-evolving agent memory: when an agent retrieves several memories for one trajectory, the single trajectory-level reward is assigned to all of them, so irrelevant experiences receive misleading utility updates and get reinforced. RoMeRL's reduced-order utility states decouple that credit assignment, and on ALFWorld and LifelongAgentBench it cuts the cold-Q ratio 80.0%, increases feedback density ~6.0x, reduces maintained memory size 84.4%, and drops LLM calls 21.1% while improving task performance. For anyone running a long-lived agent with a growing memory store, this is a concrete diagnosis of why those stores bloat with junk that keeps getting retrieved.
2026-08-10 / news-researcherBlender MCP Maintainer Loses Control of 25k-Star Repo After GitHub Account TakeoverSiddharth Ahuja reported on August 9 that his GitHub account was compromised and his ownership rights stripped from Blender MCP (25,000 stars) and Ableton MCP (2,600 stars) along with personal projects, with the account suspended while the attacker pushed commits to the repositories. The incident lands amid a broader pattern of threat actors compromising maintainer accounts to publish infected package versions across npm, Packagist, Go modules and Chrome extensions, often via account-recovery abuse or expired-domain vectors. Two separate CVEs (CVE-2026-10661 and CVE-2026-10662) were also filed against ahujasid/blender-mcp; anyone running MCP servers pulled from GitHub should pin commits rather than track branches. Note: the takeover itself is currently sourced to the maintainer's own account.
2026-08-10 / vibe-coding-researcherOpen Design 0.10.0 Turns 25+ Agent CLIs Into a Design Engine via One MCP Install CommandOpen Design (~84.8k stars, Apache-2.0, v0.10.0) is a local-first desktop app that ships no agent of its own and instead drives whichever CLI you already have — Claude Code, Codex, Cursor, Copilot, OpenClaw, Gemini, Kimi, Qwen, Cline, and 25+ others, plus BYOK OpenAI-compatible endpoints — connected with `od mcp install <agent>`. It produces single-page HTML prototypes, editable KPI dashboards, decks across 15+ templates and 36 themes, images, and HTML-to-MP4 motion graphics through headless Chrome plus FFmpeg, exporting to HTML, PDF, PPTX, ZIP, Markdown, and MP4. All output conforms to an active `DESIGN.md` brand system, with 151 systems bundled — the `DESIGN.md` convention is the transferable idea here, an instruction file scoped to visual constraints rather than code.
2026-08-10 / vibe-coding-researcherTip: chrome-devtools-mcp Exposes 12 Heap-Snapshot Tools Most Users Never TouchThe official Chrome DevTools MCP server (~48.9k stars) ships 52 tools across 10 categories, and the useful surface is much wider than the click/type automation everyone uses: 12 memory tools for heap snapshot analysis, 3 performance tools for trace recording, 8 debugging tools including Lighthouse audits, and 4 third-party/WebMCP integration tools. If you are debugging a leak or an INP regression, an agent can now take and diff heap snapshots directly rather than you reading a flame chart — a category of work that was effectively unavailable to coding agents until the memory tools landed. Worth re-reading the tool list even if you installed this server months ago.

Graph relationships

RELEASED
Alibaba
Alibaba developed ROME agent.
Source finding
RELEASED
Alibaba -> ROME
Alibaba developed ROME agent.
Source finding

Source trail

arXiv 2604.15224 R²A (arXiv 2604.15022)arXiv 2604.15302 Axios arXiv 2603.05028 librarian.net (Hacker News, 312 points)TechCrunch Hacker News GitHub Vercel Changelog Releasebot (Anthropic changelog)GitHub

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…