beacon-management-gateway

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The code implements a command-and-control beacon listener: it accepts connections from implants, performs a custom XOR/XTEA key exchange, decrypts and parses system-survey payloads (process lists, network config, uptime, etc.), handles chunked ACKed transfers, and writes the harvested data to files—behavior strongly indicative of intentional data exfiltration and a backdoor/C2 server.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime TCP listener ingests beacon payload bytes from an external connecting party (outsider) via conn.recv(...) and then parses/decrypts that data into parse_beacon_data(decrypted) / write_rsi_file(...), which becomes LLM-readable context if the agent later summarizes/uses the parsed fields.