Slash Commands

Slash commands provide special actions in HolmesGPT's interactive mode. Type any command with a leading / to execute it.

Built-in Commands

Command	Description
/help	Show all available commands and their descriptions
/exit	Exit interactive mode and return to shell
/clear	Clear conversation history and start fresh
/tools	Show available toolsets and their status
/context	Display token usage and context size information
/auto	Toggle automatic display of tool outputs
/last	Show tool outputs from the most recent AI response
/show [number\|name]	View specific tool output in scrollable modal
/run <command>	Execute shell command and optionally share with AI
/shell	Start interactive shell session

Command Details

/help

Displays all available slash commands and their descriptions.

> /help

/exit

Exits the interactive session and returns to your shell. Alternative: press Ctrl+C twice.

> /exit

/clear

Clears all conversation history and starts fresh. Useful when switching topics or when context becomes too large.

> /clear

/tools

Lists all configured toolsets and their current status (enabled/disabled).

> /tools
Available toolsets:
✓ kubernetes (enabled)
✓ prometheus (enabled)
✗ grafana (disabled - no URL configured)

/context

Displays detailed information about token usage and context size.

> /context
Context Usage:
- System: 1,250 tokens
- User: 3,421 tokens
- Assistant: 2,156 tokens
- Tools: 5,234 tokens
Total: 12,061 / 128,000 tokens (9.4%)

Top tools by token usage:
1. kubernetes_get_pod_logs: 2,341 tokens
2. kubernetes_describe_pod: 1,893 tokens

/auto

Controls whether tool outputs are automatically shown after each AI response. When disabled, use /last to view outputs.

> /auto
Auto-display tool outputs: ON

/last

Displays all tool outputs from the most recent AI response. Useful when auto-display is off or to review outputs again.

> /last

/show [number|name]

Opens a scrollable modal to view full tool output. You can specify outputs by: - Number: /show 3 (shows 3^rd tool output) - Name: /show kubernetes_get_pod_logs (shows specific tool output)

> /show 1
> /show kubernetes_get_pod_logs

Modal Navigation

• j/k or ↑/↓ - Move up/down
• g/G - Go to top/bottom
• d/u - Half page down/up
• f/b or PgDn/PgUp - Full page down/up
• w - Toggle word wrap
• q or Esc - Close modal

/run <command>

Runs a shell command and optionally shares the output with the AI.

> /run kubectl get pods -n production
# Output displayed...
Share output with AI? (Y/n): y
Add a comment or question (optional): why are some pods in CrashLoopBackOff?

/shell

Starts an interactive shell session. When you exit, you can share the entire session with the AI.

> /shell
$ kubectl logs failing-pod-xyz
$ kubectl describe pod failing-pod-xyz
$ exit
Share shell session with AI? (Y/n): y
Add a comment or question (optional): The pod keeps failing with OOM errors