overnight-agent-guard

No direct malicious payloads (e.g., exfiltration, backdoor installation, or cryptomining) are evident in this specific snippet. The dominant security risk is that the script executes AGENT_CMD via eval, enabling arbitrary command execution if AGENT_CMD can be influenced by an attacker or misconfigured CI variables. Additionally, it executes multiple helper binaries from a BIN path derived directly from NIGHTSHIFT_BIN without validation; if that environment value is untrusted, attacker-controlled binaries could run. Treat this template as unsafe unless AGENT_CMD/CHECKS and NIGHTSHIFT_BIN are strictly controlled and the invoked helper executables are trusted and integrity-checked.