linkedin-content
Audited by Socket on Mar 4, 2026
1 alert found:
MalwareThis skill is plausible and functionally consistent with its stated purpose (helping create LinkedIn posts and invoking inference.sh apps). However, it contains multiple supply-chain and data-flow risks: it recommends a pipe-to-shell installer (curl | sh), relies on binaries hosted on a custom distribution domain, and instructs transitive installs (npx skills add) that expand trust. The CLI workflow requires login and forwards user-provided content and potentially credentials to remote apps; the documentation lacks detailed, auditable descriptions of where credentials and data are sent or how they are stored. These patterns are not necessarily proof of malware, but they are high-risk supply-chain and credential-forwarding vectors. Recommend treating the installer and any transitive skill installs as untrusted until their install scripts and binaries are audited; require explicit checksum verification, prefer package-manager installs or manually inspected installers, and ensure credential storage/transmission details and privacy policies are available before use.