Introduction
The Hermes CLI auto-completes slash commands when you press Tab. That is a convenience, but it is also a discovery interface. Typing / and pressing Tab shows you every command available right now, including ones added by skills, plugins, or recent updates. Most users learn the command surface by tab-completing through it instead of reading the reference.
Key Concepts
- Tab completion: Pressing Tab after typing part of a command completes it (or lists matches).
- Prefix matching: The CLI resolves
/hto/helpwhen only one command starts withh. - Dynamic surface: Installed skills and plugins add their own slash commands. Tab completion always reflects what is loaded now.
Real World Context
A developer reads the slash commands reference once during onboarding, then never again. Two months later Hermes ships three new commands. Instead of re-reading the docs, the developer types / and Tab during their next session. The three new commands are right there in the autocomplete list. They learn the surface by living in it, not by studying it.
Deep Dive
Tab completion works at three levels:
1. From scratch. Type / and Tab. The CLI shows every slash command grouped by category. This is the cheapest way to scan what is currently available.
2. By prefix. Type /comp and Tab. The CLI either auto-completes to the only match (/compress) or lists the candidates if there is ambiguity (e.g., /comp showing /compress and /completion).
3. By argument. Some commands also tab-complete their arguments. /model followed by Tab lists configured providers and aliases. /personality followed by Tab lists installed personas. This means you can compose a full command without ever leaving the prompt.
text# Discovery from zero knowledge > /[Tab] /agents /background /branch /busy /clear /compress /config /cron /debug /help /history /image /insights /model /new /paste /personality /platforms /profile /queue ... # Refining with a prefix > /mod[Tab] /model # Argument completion > /model [Tab] anthropic/claude-sonnet-4 openai/gpt-4o-mini ollama/llama-3.3 ...
Prefix matching also works for invocation, not just completion. Hermes resolves unique prefixes even without Tab, so /h runs /help if no other command starts with h. Use this for the commands you reach for hourly; rely on Tab when prefix is ambiguous.
When a skill, plugin, or hub command adds new slash commands at runtime (e.g., after hermes skills install some-skill), they appear in the Tab list immediately without restarting the session.
Common Pitfalls
- Reading docs instead of pressing Tab: The docs lag the runtime. Tab completion never does. Beginners often miss commands that exist on their installation but not in the doc version they read.
- Stopping at the first match: When Tab shows multiple completions, scan all of them. The one you wanted may not be the alphabetically first one.
Best Practices
- Make
/+ Tab your default for the first month: It forces you to see the surface every time. After a few weeks the commands move into muscle memory. - Press Tab after every update: New commands often arrive with Hermes updates. A quick
/+ Tab is the fastest changelog scan you can do.
Summary
- Tab completion is both a typing shortcut and the canonical way to discover commands.
- It works at three levels: from scratch, by prefix, and on arguments.
- The Tab list always reflects what is currently loaded, including new skills and plugins.
- Prefer Tab over the docs for everyday command discovery; the docs lag, Tab does not.
Code Examples
# Three Tab completions, three different jobs
> /[Tab] # discover every command available right now
> /pers[Tab] # complete the personality command from a prefix
> /personality [Tab] # list the installed personas as arguments