npm detections from LPM Firewall.

This package injects non-Axios host profiling and hidden telemetry into ordinary user-initiated GET requests. The broken invocation limits successful exfiltration but does not make the im...

Source inspection confirms a concrete, unconsented data-collection and transmission path activated by normal GET usage. The absence of install hooks does not mitigate the runtime exfiltra...

The package contains a concrete, automatic metadata-exfiltration path unrelated to an HTTP client’s requested operation. The loopback destination reduces direct remote reachability but st...

Although the package has no install lifecycle hook, its explicit CLI workflow establishes persistent collection and uploads broad OpenClaw workspace content to an external package-control...

The package contains a concrete runtime exfiltration-and-remote-execution chain unrelated to its stated middleware/logger behavior. Absence of install hooks does not mitigate execution wh...

Direct source inspection confirms an unsolicited preinstall-time outbound request containing harvested local metadata. This is concrete install-time data exfiltration, not a benign packag...

Source inspection confirms unconsented postinstall writes to broad AI-agent skill roots. Policy requires blocking this concrete lifecycle control-surface mutation even though no credentia...

Direct source inspection confirms a remotely controlled arbitrary-code execution path exposed by the package runtime API. The absent lifecycle hook limits install-time behavior but does n...

The package's sole runtime entrypoint is an unsolicited full-screen remote-content loader, not a normal library interface. Although it lacks install-time behavior and local harvesting, th...

Direct source inspection confirms an automatic postinstall writes foreign Claude configuration and instruction files, then injects a mutable `@latest` MCP launcher. This meets the firewal...

This is a concrete, concealed remote-code-execution path reached by normal package use, not merely a suspicious primitive. The absence of install hooks limits automatic installation-time...

Direct source inspection confirms a concealed payload appended to the public CLI entrypoint and executed on every CLI invocation. The absence of an install hook does not mitigate this con...

Source inspection confirms an install-time mutation of foreign, broad AI-agent configuration files that persists package-controlled session instructions. This meets the blocking policy de...

Direct source inspection confirms an automatically executed, obfuscated dynamic-code loader in the public CLI entrypoint. Although it is not install-time, its concealed execution path is...

Direct source inspection confirms import-time local command execution and intentional out-of-band exfiltration to a non-package endpoint. This is concrete malicious behavior, not a benign...

Source inspection confirms concrete import-time command execution and outbound exfiltration to a fixed third-party endpoint. This is malicious behavior and not package-aligned React UI fu...

Source inspection confirms a concrete credential-exfiltration chain in a package-installed agent skill. The install hook makes the malicious capability available automatically, while exec...

Source inspection confirms an unconsented postinstall mutation of foreign, broad AI-agent control surfaces. Under the stated policy this is concrete malicious behavior regardless of the b...

Source inspection establishes intentional transmission of caller data and generated random values to unrelated endpoints with no package-aligned need. No install hook is required for this...

Direct source inspection confirms a concrete credential-exfiltration path to a non-Facebook author-controlled service, activated by the package's automatic login-recovery behavior. This e...

The package contains concrete import-time arbitrary code execution through decrypted external payloads and detached child processes. The discrepancy between source entry/build configurati...