loopany

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The content describes a remote-connected daemon that syncs local loop folders to a server, can fetch/install/update user-scoped “skills” from that server, and allows server-stored/remote-updated workflows to run locally with access to local credential files and network calls — enabling data exfiltration, supply-chain updates, and remote code execution risks if the server or a connect-key is abused.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The bootstrap and setup docs instruct the agent at runtime to fetch instructions over HTTP from the server (e.g. "/api/skill/references/create.md" and other "/api/skill/references/..." endpoints), and those fetched files are explicitly followed to drive what the agent does, so remote content fetched at runtime directly controls agent prompts/workflow.