platform-core

This skill is a documentation/definition of the AiTer platform CLI surface and agent autonomy rules. There is no direct evidence of embedded malware, obfuscated payloads, remote download-execute chains, or hardcoded credentials in the provided text. However, the combination of broad autonomous permissions (read project files, run terminal commands, send messages), ability to create tunnels, and ability to install third-party plugins creates a moderate supply-chain and exfiltration risk if the runtime environment or installed plugins are compromised. Key risks are transitive plugin installs, memory file exfiltration via messaging/webhooks or tunnels, and insufficiently gated autonomous actions. Mitigations: enforce per-action confirmations for sensitive operations, restrict plugin installation to signed/pinned sources, audit and restrict tunnel creation and messaging endpoints, and treat .aiter/memory contents as sensitive data that must not be sent without explicit user approval.