npm detections from LPM Firewall.

Source inspection confirms concrete exfiltration to an external ngrok endpoint and unconsented workspace mutation, reachable through binding.gyp build evaluation and runtime entrypoint ex...

Source inspection confirms install-time project AI-agent configuration mutation, which matches the firewall block policy even though scanner exfiltration hints were noisy. No concrete cre...

Source inspection confirms concrete install-time collection and exfiltration of host identity data to an external endpoint. This is malicious behavior, not merely suspicious network capab...

Direct source inspection confirms remote payload execution from an obfuscated endpoint on runtime activation, which is concrete malicious behavior. Lack of install hooks limits trigger ti...

Source inspection confirms unconsented npm postinstall mutation of foreign/broad AI-agent control surfaces, which meets the block policy even though the MCP functionality itself appears p...

The package crosses the LPM block boundary because postinstall unconsentedly writes package-supplied AI-agent skills/agents/control files into foreign/broad IDE directories. Scanner simil...

Under the install control surface policy, unconsented npm postinstall mutation of foreign/broad AI-agent control surfaces is blockable even when the payload is package-aligned and no exfi...

Static inspection confirms automatic postinstall mutation of multiple broad AI-agent skill directories, matching the policy for blocking unconsented install-time mutation of foreign/broad...

Source inspection confirms unconsented install-time secret harvesting and external exfiltration in postinstall.js. This is concrete malicious behavior, not package-aligned SDK functionality.

The provided scanner snippets indicate concrete install-time exfiltration behavior, but finalization requires direct source inspection and this environment did not permit further tool use...

Direct source inspection confirms unconsented install-time credential harvesting and external exfiltration unrelated to an SDK. This is concrete malicious behavior, not merely a suspiciou...

Direct source inspection confirms unconsented install-time credential and file harvesting with external exfiltration to Telegram, unrelated to an OpenSea SDK. This is concrete malicious b...

Direct source inspection confirms an install-time lifecycle script performs unconsented credential, SSH key, wallet/seed, and environment secret exfiltration to Telegram. This is concrete...

Read-only source inspection confirms concrete install-time credential and sensitive file exfiltration to Telegram with no user consent or package-aligned purpose. This is malicious malwar...

Direct source inspection confirms install-time credential and sensitive-file exfiltration unrelated to the advertised web3 SDK. This is concrete malicious behavior with automatic lifecycl...

Direct source inspection confirms unconsented install-time credential harvesting and exfiltration to a hardcoded Telegram bot. This is concrete malicious behavior unrelated to a Polygon SDK.

Static inspection confirms a postinstall hook mutates broad/foreign AI-agent control surfaces without requiring an explicit setup command, which matches the firewall block policy. The res...

Static inspection confirms a postinstall external sync command tied to AI-agent hook/control-surface behavior, which meets the policy boundary for blocking unconsented install-time mutati...

The decisive behavior is not ordinary CLI networking; it is unconsented postinstall mutation of foreign/broad AI-agent control surfaces. Under the provided policy this should be blocked e...

Static inspection confirms unconsented postinstall mutation of OpenCode global config and AGENTS.md, which meets the firewall block policy for AI-agent control-surface writes. Other risky...

Source inspection confirms install-time mutation of broad AI-agent control surfaces through a postinstall hook, which meets the block policy. The scanner finding is substantiated by packa...