Sonatype
Buying intent
84 tracked signals | Top 15 topics are below | Security and Sales are carrying most of it.
Attention by team
LinkedIn activity, by teamWhere Sonatype's own people are actually spending their attention, by team, by topic. Bands run Low to High against the busiest pairing on this page, and each cell also shows how much of that team's own activity it represents.
Topics being researched
30-day windowEvery tracked topic, ranked by volume, not by our guess at what matters. Confidence is the classifier's own certainty that a signal belongs where we've filed it.
10d ago
10d ago
18d ago
11d ago
18d ago
11d ago
18d ago
12d ago
24d ago
24d ago
24d ago
16d ago
25d ago
22d ago
26d ago
Need intent for a specific topic or industry?
We track the full taxonomy across every account in the graph — including themes not shown on this page.
Who's active at Sonatype
verified title on fileTitles, seniority and topic straight from each person's own activity, with a LinkedIn link so you can check any of them yourself.
See everyone, not just the first 10
21 people across every department at Sonatype, plus a LinkedIn profile link for each.
Primary products / business lines
LinkedIn company profileThe Sonatype journey started 15 years ago, just as the concept of “open source” software development was gaining steam. From our humble beginning as core contributors to Apache Maven, to supporting the world’s largest repository of open source components (Central), to distributing the world's most popular repository manager (Sonatype Nexus Repository), we’ve played a meaningful role in helping the
New capability sought
Employee posts (LinkedIn)Supply Chain Security; AI Agent Software; AI Software; Agentic AI System; Intellectual Property; Product Ownership; Software Development
Top accounts researching Sonatype
names withheld on the public pageThese are companies whose own people brought up Sonatype unprompted, not accounts we guessed might be interested. We can't yet tell an implementation partner from a genuine buyer here, names unlock along with the buyer profile below.
- Account withheldGas TSO of Ukraine1 contacts
- Account withheldProject Eleven1 contacts
- Account withheldFreshworks1 contacts
Buyer profile
company size · seniorityCompany size and how senior the people involved are, the two things that decide whether this is a real deal. Competitor overlap isn't computed yet for this account.
Buying committee functions
Employee job titles (LinkedIn)Engineering — 5 people; Sales — 3 people; Leadership — 2 people; HR / Talent — 2 people; Marketing — 1 person; Operations — 1 person; Security — 1 person; Customer Success — 1 person
What's been said
public posts by Sonatype's teamNo public post naming Sonatype has surfaced in the past year, so this is what Sonatype's own team is posting about publicly — their topics, in their words.
We found a high-severity access control bypass in one of the more popular MCP servers in the Kubernetes ecosystem (CVE-2026-46519, CVSS 8.8), one with 20K+ weekly npm downloads and 1.4K GitHub stars. The project ships 3 environment variables to restrict what agents can do, including a "read-only mode." All 3 are enforced at the discovery layer but *not the execution layer*. The agent asks the server what tools are available and the restricted ones don't show up. But any client can invoke them directly. And the tool names? They're in the README. In the worst case, this is full cluster compromise for anyone who can reach the endpoint. Fixed in v3.6.0. If you're running an earlier version, update now. Credit to Francisco Rosales of Manifold Security for finding and reporting this one. Shoutout to the project maintainer Suyog Sonwalkar for acknowledging, fixing, and releasing same-day - exemplary vulnerability management! 🫡 Link to full disclosure in the comments 👇 cc Neal Swaelens Oleksandr Yaremchuk
May 2026Grafana's transparency here is genuinely good. A public disclosure, no ransom paid, FBI stance cited. Respect 🫡 The replies, however, ask the obvious question: if it's open source, what exactly did the attacker get?? The answer matters more than people realise. OSS repos often have private forks, internal branches, unreleased features, or proprietary config/infra code sitting alongside the public stuff. A GitHub token with broad org access goes far beyond what you see on grafana[.]com/oss. So, the real story here isn't "they stole the code" but *what all* that token had access to. Credentials with over-privileged #GitHub access are incredibly common, and most orgs wouldn't know what to do with the blast radius analysis even if they tried. Grafana Labs handled this well. Look forward to what their post-incident review surfaces about token hygiene (and whether they share it publicly). That'd be the actually useful disclosure.
May 2026Heard of 'lethal trifecta'? Simon Willison coined it. A CISO's article this week got me thinking. If every legitimate AI agent hits all 3 criteria... what's the signal now?
May 2026One widely used enterprise scanner flagged more than 40% as potentially malicious, when we scanned 19,000+ skills across public registries. But most were actually fine. The skills flagged included popular productivity tools and widely deployed coding assistants. They were flagged because their instructions contained imperative language, references to system access, conditional logic. All normal features of a functional skill. At that alert volume, security teams either tune the tooling so aggressively it stops catching real threats, or they ignore it. Neither is a useful security posture. Meanwhile, a skill called "derp" passed two of three scanners clean. What it actually does: instructs the agent to *produce broken code* on every task, silently, with professional confidence. If the developer gets suspicious, the skill tells the agent to blame their environment. The scanners saw natural language instructions and found nothing wrong. The harm only exists at runtime, in what the agent does after loading a skill: - which files it reads, - which APIs it calls, - which credentials it touches, - which subprocesses it spawns. This is the same gap signature-based antivirus hit against living-off-the-land attacks, except agents have broader access than any traditional endpoint ever did. Full analysis in the comments.
May 2026🚨 XBOW has had early access to #Mythos Preview for the last month! As advertised, It's a major step forward when it comes to LLM-based security. That said, it's still a brain that requires a body, like what we've built here at XBOW . How'd Mythos perform? Read about it here: https://lnkd.in/e3G8XuZZ
May 2026Attackers are abusing Google ads + Claude[.]ai shared chats to push macOS malware. When you google "claude mac download" (I just tried), a sponsored search result points to Anthropic's real domain, so there's no fake URL to spot. Berk A. found an active Claude[.]ai chat leading users to malware. While reproducing Albayrak's finding, I discovered a *second* active chat leading users to a different variant! https://lnkd.in/en-GCKDK Always go to the official source, such as claude[.]ai yourself to download native apps. When the real URL itself is the threat, the usual advice to "check where the link goes" won't cut it.
May 2026