npm security reports, with evidence.

OpenSSF/OSV malware advisory MAL-2026-15569 blocks this version. LPM AI assessed this version as suspicious; the assessments disagree. dist/client.js executes initializeProvider() at modu...

OpenSSF/OSV malware advisory MAL-2026-15569 blocks this version. LPM AI assessed this version as suspicious; the assessments disagree. dist/client.js executes initializeProvider() at modu...

OpenSSF/OSV malware advisory MAL-2026-15567 blocks this version. LPM AI assessed this version as suspicious; the assessments disagree. Package is published as htps-provider with a README...

OpenSSF/OSV malware advisory MAL-2026-14562 blocks this version. morglog@1.1.0 is published under a name resembling the popular morgan HTTP logger and reuses morgan's description. The tar...

OpenSSF/OSV malware advisory MAL-2026-14561 blocks this version. The package presents itself as `relative-import-path` (README title, homepage and repository URL point at github.com/marko...

This is concrete install-hook abuse: an npm postinstall mutates the foreign Codex control surface and executes remote shell content. The relay-driven bypass retry materially increases the...

The package has a concrete automatic install-time persistence mechanism that mutates a foreign project and causes ongoing external reporting. The absence of direct token transmission does...

The package embeds a live remote command channel that executes operator-provided JavaScript and returns output. No install hook is present, but the runtime behavior is concrete malicious...

The automatic postinstall writes package-controlled instructions into several foreign AI-agent skill locations. This meets the policy threshold for malicious install-hook abuse even thoug...

The package implements concrete persistent agent-data collection and upload behavior, plus a latest-version self-install path and an automatic external binary download. The explicit hook-...

The package implements automatic disclosure of sensitive recruitment and conversation data to a hard-coded third-party endpoint. The lack of an install hook limits scope but does not remo...

The package combines an automatic lifecycle hook, broad host-agent configuration writes, injected agent instructions, and runtime self-updating through a fresh global install. The global-...

The automatic lifecycle hook changes foreign AI-agent configuration and introduces a moving remote command path. This meets the install-control-surface blocking policy.

The postinstall lifecycle performs unconsented broad mutation of foreign AI-agent control surfaces. This meets the install-control-surface blocking policy despite no confirmed installer e...

The package combines an automatic install hook, writes to multiple foreign AI-agent configuration surfaces, and injected instructions that direct agents to use its relay without user appr...

This is a concrete, unconsented postinstall modification of a foreign GitHub Actions control surface with network behavior. It meets the install-hook abuse blocking policy.

The automatic postinstall chain performs remote code installation and unconsented Codex plugin configuration. Its global-install guard narrows exposure but does not establish user consent...

The package deliberately conceals a remote code loader in its runtime entrypoint. Its network-selected payload execution is concrete malicious behavior.

This is an unconsented postinstall mutation of broad AI-agent control surfaces combined with automatic telemetry. The behavior is concrete and reaches user-wide configuration and arbitrar...

The package contains a concealed import-time loader that obtains and executes remote code, then persists it as a detached process. The absence of lifecycle hooks does not mitigate this di...

The automatic postinstall and normal startup paths modify dependencies and the host system, including privileged package-manager attempts. The package also automatically maintains hooks i...