New on the FIRST blog: Jenn Gile, Co-Founder of OpenSourceMalware and #VulnCon26 speaker, on why malicious open source packages don't fit the traditional vulnerability intelligence model.
The response motion looks familiar. A malicious package appears in a public registry, a record lands in OSV, tools fire an alert, and someone opens a ticket. But the data and the playbook don't actually match the threat.
🔍 Vulnerabilities are passive. They wait to be exploited.
⚡ Malicious packages are active. They execute on install.
🔧 Vulnerabilities have a fixed version.
🚫 Malicious packages ARE the latest version.
That mismatch leaves three investigative gaps vulnerability databases weren't built to fill:
📦 Payload: what the malware did and which files were affected.
👤 Threat actor: C2 infrastructure and accounts reused across campaigns.
🔗 Campaign: how one package connects to broader activity.
Case in point: the axios account takeover on March 30, 2026. OSV surfaces three IOCs. The campaign has at least nine, two of them shared with other malicious assets.
Jenn's argument: malicious packages need their own intelligence track, built around a different set of questions.
📖 Read more: https://go.first.org/BwFfv
Malicious Packages Don't Fit the Vulnerability Intelligence Model
Malicious open source packages and software vulnerabilities may look alike on the surface, but they demand entirely different response playbooks. Treating a malicious npm or PyPI package like a CVE leaves critical questions unanswered: what did it execute, where did it phone home, and what campaign is it part of? Purpose-built malicious package intelligence infrastructure is needed to answer those questions.
