Scylla brings supported terminal and SSH activity into an inspectable workflow. Local commands run with the permissions of the executing Windows account. Remote commands run through a configured connection with the permissions of the remote account.
Before execution
- Confirm the active project and working directory.
- Review the effective Security Policy for terminal or SSH operations.
- Configure the connection and protected credential references in Scylla Keyring.
- Start with a read-only command when checking a new system.
Review the result
Monitor job state and inspect the filtered output returned to the agent. Protected credentials should stay in the execution path rather than the conversation. For production systems, verify the target connection and requested command before approving a write.