hcom-agent-messaging

[Skill Scanner] Installation of third-party script detected All findings: [HIGH] supply_chain: Installation of third-party script detected (SC006) [AITech 9.1.4] [HIGH] supply_chain: Installation of third-party script detected (SC006) [AITech 9.1.4] No explicit malicious code is present in the provided documentation, but the skill requests and enables high-privilege operations (terminal screen reading, input injection into other agent processes, spawn/kill agents, and cross-device MQTT relays) and instructs a pip install + auto-install hooks pattern. These behaviors are disproportionate to a low-privilege helper and present substantial supply-chain and exfiltration risk if the package or relay are compromised. Treat this skill as suspicious: audit the actual package code, verify install sources and signatures, review hook scripts, restrict relay endpoints (self-host where possible), and require explicit user approval for any terminal injection or process control actions. LLM verification: The SKILL.md describes a high-privilege orchestration tool that matches its stated purpose (agent messaging and coordination) but exposes a number of supply-chain and privilege risks. Key concerns: unpinned pip installs (supply-chain risk), hooks/daemon/relay auto-install with opaque behavior, terminal-screen reading and input injection (high privilege and potential for misuse), and unspecified network/MQTT relay endpoints (possible exfiltration). No direct evidence of embedded malware in the do