Slash Commands Reference

Slash commands are entered directly in the prompt composer starting with /. Pressing / at the start of a prompt opens an interactive completion menu.

Command Index

Command	Arguments	Description
/model	[model-name]	Change the active model (shortcut: Ctrl+O).
/effort	[level]	Set reasoning effort (default, none, minimal, low, medium, high, xhigh) (shortcut: Ctrl+T).
/plan	—	Toggle read-only Plan mode (shortcut: Shift+Tab).
/lite	—	Toggle lightweight terminal mode (minimal prompt, session-only context).
/mentor	—	Toggle collaborative Mentor mode with dual-model consultation.
/orchestrator	—	Toggle prompt-guided Orchestrator mode (delegates all tool-backed work).
/profile	[name]	Switch the active operating profile (standard, plan, lite, mentor, orchestrator).
/auto-approve	[off|advisory|auto|always]	Set shell and tool auto-approval policy. always disables sandbox.enabled.
/sandbox	—	Toggle shell sandbox isolation on or off.
/compact	—	Compact older conversation turns into a local summary.
/rewind	[last|<turn>] [edit|resend]	Rewind conversation history with interactive discard inspection.
/undo	—	Alias for /rewind edit (places target turn back in composer for revision).
/retry	—	Alias for /rewind resend (immediately resends the last user turn).
/retry-turn	—	Retry the last turn that the provider failed to complete (e.g. after network errors).
/retry-tool	—	Re-execute the last failed or timed-out tool call.
/copy	[N]	Copy the latest assistant response (or $N$-th prior response) to clipboard.
/usage	—	Show token usage and model cost metrics for the current session.
/resume	[conversation-id|ls]	Resume a saved conversation or browse conversations.
/handoff	—	Hand off the last assistant response to another model.
/providers	—	Open interactive provider manager (list, add, edit, remove providers or switch OAuth accounts).
/skills	[skill-name]	Activate a skill for the next request.
/settings	[key] [value]	View or modify runtime configuration settings.
/clear	—	Clear current conversation history and start a fresh session.
/quit	—	Exit term2 session cleanly.

---

Detailed Command Documentation

/model [model-name]

• Usage: /model, /model gpt-5.1, /model openrouter/anthropic/claude-3.7-sonnet
• Shortcut: Ctrl+O
• Switches the primary model. If called without arguments, opens the interactive model picker menu.

/effort [level]

• Usage: /effort high, /effort medium, /effort none
• Shortcut: Ctrl+T
• Sets the reasoning effort level passed to models supporting variable reasoning tokens.

/profile [profile-name]

• Usage: /profile, /profile standard, /profile plan, /profile lite, /profile mentor, /profile orchestrator
• Switches the active operating profile. Switching to or from lite mid-session prompts for confirmation or recommends /clear.

/auto-approve [mode]

• Usage: /auto-approve off, /auto-approve advisory, /auto-approve auto, /auto-approve always
• Sets the tool execution approval mode. Mode always sets sandbox.enabled = false.

/rewind [last|<turn>] [edit|resend]

• Usage: /rewind, /rewind 3 edit, /rewind last resend
• Opens the rewind menu to select a turn to rewind. Discards subsequent transcript turns and either loads the user prompt into the composer (edit) or resubmits it immediately (resend).

/retry-turn

• Usage: /retry-turn
• Retries the last assistant turn when a provider error, network timeout, or rate limit prevented the response from completing.

/retry-tool

• Usage: /retry-tool
• Re-executes only the last failed or timed-out tool call without losing conversation context or discarding subsequent messages.

/copy [N]

• Usage: /copy, /copy 2
• Copies the latest assistant response text (or the $N$-th prior response) to your system clipboard using OSC 52 or platform clipboard utilities. If multiple copyable sections exist, opens a selection menu.

/handoff

• Usage: /handoff
• Captures and passes the latest assistant response to the interactive handoff flow.