Fetching from the wire…
Public story · 2026-07-21 · high
The listing bets describing sites beats driving them, the approach that earned browser-use 105,000 stars.
Why now: This surfaced in agent-tooling coverage on July 21, with the describe-versus-drive debate for reading the web still open.
Manifest launched on Product Hunt with a pitch to turn any webpage into an agent-executable action manifest, per the listing.
The stakes are how agents act for users without clicking through a UI, and browser-use already has 105,000 GitHub stars doing it the opposite way.
Manifest's pitch is to solve that by describing the site instead of driving it.
Which method wins is genuinely unsettled. Manifest needs a site to publish its own manifest before an agent can use it. Browser-use attacks the same problem from the driving side, needing no cooperation from the site at all.
Product Hunt's page doesn't include independent verification of Manifest's capability claims, so the demo reads as a pitch, not a benchmark.
My take: driving beats describing, at least for now. Browser-use already works on sites that never opted in. Most sites have no reason to maintain a manifest file for traffic they didn't ask for. Manifest's approach only pays off if enough sites decide to publish one, and nothing in the launch says that's happening yet.
This surfaced in agent-tooling coverage on July 21, with the describe-versus-drive debate for reading the web still open.
Each link below shares sources, entities, or timing with this story.
If you have a CLAUDE.md, you're in scope. Today. arXiv 2607.14611 (cs.CR, filed July 16) evaluates prompt injection planted in the persistent memory files that agentic coding systems write and re-read across sessions. The researchers tested both Anthropic's Claude Code and Ope...
August 15's board leads with Inferock Bench at 154 upvotes, a local open-source proxy that sits between your app and OpenAI, Anthropic, Gemini or OpenRouter-shaped calls and emits a per-call billing-integrity receipt. It catches streams cut mid-sentence that were still billed,...
Spotify's Portal team published Xirp on August 10: a vendor-neutral agentic development environment that manages concurrent sessions across Claude Code, Gemini CLI, and Codex, each session isolated in its own git worktree so dozens of agents can work the same codebase without...
A joint study across OpenAI, Anthropic, and Google DeepMind found no single filter or classifier holds up against an adaptive attacker. (Help Net Security) Treat injection as a containment problem, not a detection one: separate trusted from untrusted text, validate output stru...
I check Product Hunt maybe once a week and usually regret it. Today's board is worth reading as market structure. The July 30 leaderboard: SKI at 277 upvotes (free voice input for Claude Code and Codex). AI Search Console at 249 (prompt analytics and citation mapping). Memmy A...
This one's been building for days and it crystallized this week. Per The Register, the incident behind the US export-control block on Anthropic's Fable 5 and Mythos 5 wasn't a jailbreak or a guardrail bypass. It was a plain three-word prompt, "fix this code," run against CVE-l...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.