hermes-agent-spawning

To spawn autonomous agents for long-running tasks, you launch independent Hermes CLI processes that run in isolation with their own sessions, tools, and environments. This improves reliability and scalability for complex workflows.

Spawning autonomous agent subprocesses requires the hermes CLI installed on your PATH and API keys configured in the ~/.hermes/.env file. These prerequisites ensure spawned processes can authenticate and operate independently.

Yes, you can run parallel workflows by spawning multiple agents concurrently. The system launches independent processes to execute parallel tasks across different projects, enabling multi-agent coordination for backend automation.

Yes, PTY-mode interactive collaboration supports back-and-forth prompts for steering autonomous agents. You can interactively guide persistent sessions that resume previous CLI history via SQLite memory.

One-shot non-interactive tasks execute a single command and exit using the -q flag, while persistent interactive sessions maintain a PTY connection for back-and-forth collaboration. Both modes spawn isolated agent processes.

Session persistence works by resuming previous CLI sessions via SQLite memory and history. This allows spawned autonomous agents to retain context across interruptions, supporting long-running research and background automation tasks.