Verified
New since your last visit
Unverified
Scope: Hermes Agent is a general autonomous agent, and this placemat covers only its CLI and TUI coding surface. Everything outside that surface is deliberately left out — the full documentation covers it.
TUI & Symbols
Keybindings
| Item | Description |
|---|---|
#Ctrl+P | Open the fuzzy command palette
|
#Ctrl+T | Expand the live-subagent dock into the full /agents roster
|
#Ctrl+X | Open the live session switcher
|
#Ctrl+G / Ctrl+X Ctrl+E | Open the input buffer in $EDITOR for long prompts
|
#Ctrl+V | Paste text, falling back to clipboard image attach
|
Transcript symbols
| Item | Description |
|---|---|
#❯ | Input prompt — where you type
|
#● | A tool call; the bullet precedes the tool name and arguments
|
#▸ / ▾ | Collapsed or expanded section chevron
|
#▍ | Streaming cursor — blinks while the model is still emitting text |
#│ / ├─ / └─ | Tree rails connecting a parent to its children
|
Status bar
| Item | Description |
|---|---|
#⏱ / ⏲ | Per-prompt elapsed time while the turn runs
|
#cmp N / ▶ N / ⛓ N / N bg | Session counters
|
#⚠ YOLO | YOLO mode is on — dangerous commands auto-approve |
#◉ focus | Focus view is on (reduced output)
|
Overlays & prompts
| Item | Description |
|---|---|
#⚠ approval required | A tool needs your explicit yes
|
#○ / ■ / ⌛ | Subagent states in the /agents overlay
|
#✎ draft | That session’s composer holds an unsent draft
|
Slash Commands (Core)
Session
| Item | Description |
|---|---|
#/new / /reset | Start a new session with a fresh id and history
|
#/clear | Clear the screen and start a new session |
#/resume | Resume a previously-named session |
#/sessions / /switch | Browse and resume previous sessions
|
#/branch / /fork | Branch the current session to explore a different path |
#/history / /save | Show conversation history, or save the conversation |
#/title | Set a title for the current session |
#/undo / /retry | Remove the last exchange, or resend the last message |
#/prompt / /compose | Compose the next prompt in $EDITOR as markdown |
#/quit / /exit | Exit the CLI |
Context & history
| Item | Description |
|---|---|
#/compress | Manually compress context — flush memories and summarise
|
#/context / /ctx | Visual context-window breakdown
|
#/status | Session info plus a locally-computed session recap
|
#/usage | Token usage, cost breakdown and session duration
|
Coding workflow
| Item | Description |
|---|---|
#/diff | Show git changes in the working directory
|
#/rollback | List or restore filesystem checkpoints |
#/snapshot / /snap | Create or restore snapshots of Hermes config and state
|
#/worktree | Inspect or create isolated git worktrees mid-session
|
#/plan | Write a markdown implementation plan to .hermes/plans/
|
#/init | Generate or update AGENTS.md from a repo scan
|
#/review | Spawn an independent full-privilege reviewer subagent
|
Run control
| Item | Description |
|---|---|
#/queue / /q | Queue a prompt for the next turn without interrupting
|
#/steer | Inject a mid-run note delivered after the next tool call
|
#/stop | Kill all running background processes |
#/bg | Run a prompt in a separate background session
|
#/btw | Ask a quick side question about the current conversation
|
#/agents / /tasks | Show active agents and running tasks |
#/moa | Run one prompt through the default Mixture of Agents preset
|
#/busy | Control what happens when you message mid-turn
|
Slash Commands (Tools)
Model & reasoning
| Item | Description |
|---|---|
#/model | Show or change the current model
|
#/reasoning | Manage reasoning effort and display
|
#/fast | Toggle provider priority processing
|
#/codex-runtime | Toggle the optional Codex app-server runtime for OpenAI models
|
#/personality | Set a predefined personality; none clears the overlay |
Tools & skills
| Item | Description |
|---|---|
#/tools | List, disable or enable tools for this session
|
#/toolsets | List available toolsets |
#/skills | Search, install, inspect and manage skills
|
#/bundles | List skill bundles — /<name> aliases that preload several skills |
#/learn | Distil a reusable skill from a directory, URL, workflow or notes
|
#/memory | Review pending memory writes and toggle the write gate |
#/refine | Run the background memory and skill self-improvement review now
|
#/reload-skills | Re-scan the skills directory for installed or removed skills |
#/reload-mcp | Reload MCP servers from config.yaml and re-probe tools |
#/reload | Reload .env into the running session
|
#/plugins | List installed plugins and their status |
#/browser | Attach browser tools to a running Chromium-family browser over CDP
|
Display & safety
| Item | Description |
|---|---|
#/yolo | Toggle YOLO mode — skip all dangerous-command approval prompts |
#/approvals | Show or set the persistent approval mode
|
#/focus | Reduced-output view showing only your prompt and the final response
|
Profile & account
| Item | Description |
|---|---|
#/profile | Show the active profile name and home directory |
#/export | Pack a profile into a shareable archive
|
#/import | Install a profile archive as a new profile
|
#/login | Sign in with a Nous account
|
#/help | Show available commands, grouped by category
|
#/version | Show version, build and environment info |
#/update | Update Hermes Agent to the latest version |
#/debug | Upload a debug report and get shareable links |
CLI & Subcommands
Global options
| Item | Description |
|---|---|
#--version, -V | Show version and exit |
#--profile <name>, -p | Select the profile for this invocation
|
#--resume <session>, -r | Resume a session by id or title
|
#--continue [name], -c | Resume the most recent session, or the most recent matching a title |
#--in <dir> | Change into a directory before starting or resuming
|
#--worktree, -w | Start in an isolated git worktree for parallel-agent work |
#--yolo | Bypass dangerous-command approval prompts |
#--ignore-user-config | Ignore ~/.hermes/config.yaml and use built-in defaults
|
#--tui / --cli | Launch the TUI, or force the classic REPL
|
#--pass-session-id | Include the session id in the agent’s system prompt |
Chat & models
| Item | Description |
|---|---|
#hermes chat | Interactive or one-shot chat with the agent |
#hermes -z <prompt> | Scripted one-shot run |
#hermes model | Interactively choose the default provider and model |
#hermes moa | Configure named Mixture of Agents presets |
#hermes fallback | Manage fallback providers tried when the primary model errors |
#hermes lsp | Manage LSP integration for semantic diagnostics on writes |
#--format stream-json | Emit structured JSONL events for a scripted hermes chat -q run
|
Sessions & projects
| Item | Description |
|---|---|
#hermes sessions | Browse, export, prune, rename and delete sessions
|
#hermes project | Manage named multi-folder workspaces
|
#hermes checkpoints | Inspect, prune or clear the shadow store behind /rollback
|
#hermes profile | Manage profiles — multiple isolated Hermes instances |
Config & auth
| Item | Description |
|---|---|
#hermes config | Show, edit, migrate and query configuration
|
#hermes setup | Interactive setup wizard for all or part of the configuration |
#hermes auth | Add, list, remove, reset and check credentials
|
#hermes migrate | Rewrite config.yaml to replace retired models or deprecated settings |
#hermes status | Show agent and auth status |
Extensibility
| Item | Description |
|---|---|
#hermes skills | Browse, install, publish, audit and configure skills |
#hermes bundles | Group several skills under one /<name> slash command |
#hermes plugins | Install, enable, disable and remove plugins |
#hermes mcp | Manage MCP server configurations, and run Hermes as an MCP server |
#hermes acp | Run Hermes as an ACP server for editor integration |
#hermes tools | Configure enabled tools per platform |
#hermes hooks | Inspect, approve or remove shell-script hooks declared in config.yaml |
#hermes import-agent | Import a Claude Code or Codex CLI setup |
Automation
| Item | Description |
|---|---|
#hermes cron | Inspect and tick the cron scheduler |
#hermes kanban | Multi-profile collaboration board — tasks, links, dispatcher |
#hermes gateway | Run or manage the messaging gateway service
|
#hermes portal | Nous Portal status and Tool Gateway routing |
Diagnostics & maintenance
| Item | Description |
|---|---|
#hermes doctor | Diagnose config and dependency issues |
#hermes dump / hermes debug | Copy-pasteable setup summary for support |
#hermes prompt-size | Byte breakdown of the system prompt and tool schemas
|
#hermes logs | View, tail and filter agent, gateway and error logs |
#hermes security audit | On-demand supply-chain audit for the venv, plugins and pinned MCP servers
|
#hermes backup / hermes import | Back up the Hermes home directory to a zip
|
#hermes completion | Print shell completion scripts (bash, zsh, fish) |
#hermes update | Pull the latest code and reinstall dependencies
|
Settings (config.yaml)
Files & precedence
| Item | Description |
|---|---|
#~/.hermes/config.yaml | Primary config for every non-secret setting |
#~/.hermes/.env / ~/.hermes/auth.json | API keys and secrets; required for credentials
|
#hermes config set | Route a value to the right file automatically
|
#${VAR} | Environment-variable substitution inside config.yaml |
Terminal backends
| Item | Description |
|---|---|
#terminal.backend | Where the agent’s shell commands run
|
#terminal.timeout | Per-command timeout in seconds |
#terminal.temp_dir | Session temp root for background logs, code sandboxes and spilled results
|
#terminal.env_passthrough | Env var names forwarded into sandboxed execution |
#terminal.home_mode | Subprocess HOME policy: auto, real or profile |
Agent behaviour
| Item | Description |
|---|---|
#model.default / model.provider | Default model and provider |
#model.persist_switch_by_default | Whether a plain /model <name> persists to config |
#auxiliary.review | Dedicated model for the /review reviewer subagent |
#goals.max_turns | Turn budget for a standing /goal; defaults to 20 |
#agent.fast_auto_seconds | Window used by /fast auto; defaults to 60 s |
#worktree_sync | How /worktree new and hermes -w branch from the remote tip |
#display.interface | Classic CLI or TUI by default
|
#display.vim_mode | CLI-only vi/vim keybindings for the input composer |
Context & memory
| Item | Description |
|---|---|
#compression | Context compression thresholds and behaviour |
#context | Context engine, pressure warnings and file truncation limits |
#database | Database settings for sessions and state |
#updates | Update behaviour |
Limits & safety
| Item | Description |
|---|---|
#iteration budget / wall-clock run budget | Caps on how long a single run may go
|
#verify-on-stop | Coding verification before a turn is allowed to end
|
Environment Variables
Locations & interface
| Item | Description |
|---|---|
#HERMES_HOME | Hermes home directory; defaults to ~/.hermes
|
#HERMES_TUI | Launch the TUI; equivalent to --tui |
Providers & tools
| Item | Description |
|---|---|
#~/.hermes/.env | Where provider API keys and tokens belong
|
Backends
| Item | Description |
|---|---|
#terminal backend variables | Credentials and endpoints for Docker, SSH, Modal, Daytona and Singularity
|
Diagnostics
| Item | Description |
|---|---|
#hermes logs | Log files live under ~/.hermes/logs/
|
Skills, Context & Subagents
Context files
| Item | Description |
|---|---|
#.hermes.md / HERMES.md | Project instructions, highest priority
|
#AGENTS.md | Project instructions, conventions and architecture
|
#AGENTS.override.md | Personal per-directory override, usually gitignored
|
#CLAUDE.md | Claude Code context files are detected too |
#.cursorrules / .cursor/rules/*.mdc | Cursor IDE coding conventions, from the working directory only |
#SOUL.md | Global personality and tone for this Hermes instance
|
#first match wins | Only one project context type loads per session
|
#progressive discovery | Subdirectory context files are injected as the agent navigates into them |
#--ignore-rules | Skip auto-injection of context files, memory and preloaded skills |
Memory
| Item | Description |
|---|---|
#~/.hermes/memories/ | Persistent memory store
|
#memory.write_approval | Stage memory writes for review before they land
|
#hermes memory | Configure the external memory provider |
Skills
| Item | Description |
|---|---|
#~/.hermes/skills/ | Agent-created skills, managed by the skill tool |
#SKILL.md | The file a skill is authored in
|
#skills.write_approval | Gate agent skill writes behind review
|
#skills.auto_load | Pin named skills to load into every new session's prompt
|
Automation & MCP
Scheduled & recurring
| Item | Description |
|---|---|
#/cron | Manage scheduled tasks from chat
|
#/heartbeat / /hb | A recurring prompt that re-enters this session when idle
|
#/loop / /proactive | Re-run a prompt on a recurring interval within this session
|
Goals & loops
| Item | Description |
|---|---|
#/goal | A standing goal Hermes works toward across turns
|
#/subgoal | Append a criterion to the active goal mid-loop
|
Suggested & templated automation
| Item | Description |
|---|---|
#/suggestions / /suggest | Review suggested automations
|
#/blueprint / /bp | Set up an automation from a template
|
Kanban & projects
| Item | Description |
|---|---|
#/kanban | Drive the multi-profile collaboration board from chat
|
No matches — clear search to see everything.