What CLI is for
The CLI is useful when a workflow already involves files, repositories, or scripts and you want to see the agent's tool activity. Official documentation covers interactive sessions, single-query mode, session continuation, and toolset selection. Start from the correct directory and profile so the agent sees the intended project context. For repeatable work, preserve the prompt and input boundaries rather than relying on a long conversation to supply hidden requirements.
What you need
- A supported Hermes installation available to the shell and a configured model route.
- The intended working directory, profile, project instructions, and permissions for the tools being requested.
Set it up
- 1
Check the installed CLI's help and current environment before adapting a command from another release or operating system.
- 2
Run a small read-only task with explicit inputs and output expectations. Use a prompt file for lengthy instructions where supported, and retain the command and resulting artifact together.
Check it works
- Inspect the actual saved file or retrieved record, and confirm which model, profile, and working directory were used.
- When resuming, check the project context and outstanding task before allowing further changes.
Limits and failure points
- A completed terminal process can still produce an incomplete answer or fail to deliver an artifact.
- Non-interactive work needs explicit failure behavior because nobody may be present to answer prompts. Shell quoting and platform-specific paths require care.
Primary sources
Spotted something wrong or out of date? Tell us on X at @HermesWatcher (opens in a new tab) and we will fix it.