Slash Commands

Slash commands are entered inside the Inferoa TUI. They form a small, purpose-built command registry — they are not a mirror of legacy CLI subcommands.

Top-Level Commands

Command	Purpose
/setup	Open endpoint, provider, web search, and Omni setup wizard
/model	Open model and provider selector
/system	Show model, web search, Omni, and runtime status (also /endpoint, /endpoints)
/access	Change this workspace's file and tool access
/skills	List discovered skills or manage enabled skills
/loop	Start or manage task and research loops
/inbox	Show loop review, blocker, automation, and self-improve items
/self-improve	Manage loop learning proposals and replay reports
/plan	Start or manage plan mode
/tokenmaxxing	Show token, cache, RTK, and routing savings
/context	Show context usage, compression state, and code intelligence
/compact	Compact conversation history into recoverable memory; optional instructions may follow
/tools	Show fixed tool schemas and renderer status
/sessions	Manage chat sessions
/daemon	Manage background daemon runs
/doctor	Check endpoint, tool, and optional Omni health
/help	Show keyboard shortcuts and slash commands
/clear	Start a fresh session
/resume	Resume a previous session
/exit	Exit the TUI

Common Subcommands

# Loop mode
/loop status                 Show active loop state
/loop                        Start a recursive long-horizon loop
/loop run deliver            Start an end-to-end delivery loop
/loop run discover           Start a metric-driven discovery loop
/loop run replay             Repeat one prompt for a fixed count
/loop run deliver --at-least 24h
                             Keep a delivery loop running for at least 24h
/loop run replay --count 100 Repeat the prompt 100 times
/loop review                 Review a pending loop decision
/loop verify                 Run an independent loop verification pass
/loop pause                  Pause the current loop
/loop resume                 Resume a paused loop
/loop drop                   Drop the current loop

# Self-improve
/self-improve help           Show self-improve workflow and commands
/self-improve status         Show learning evidence and proposals
/self-improve propose        Stage a learned workspace skill proposal
/self-improve run --replay   Run structured replay/gating
/self-improve report         Show the latest replay report
/self-improve adopt          Adopt a staged learned workspace skill

# Plan mode
/plan show                   Show active plan state
/plan set                    Set or replace the plan objective
/plan pause                  Pause the current plan
/plan resume                 Resume a paused plan
/plan approve                Approve the current plan for execution
/plan drop                   Drop the current plan

# Tokenmaxxing
/tokenmaxxing                 Show token, cache, RTK, and routing savings
/tokenmaxxing trend           Show pageable metric trend panels
/tokenmaxxing signals         Show raw lifecycle and endpoint evidence rows

# Skills
/skills list                 Show discovered skills
/skills manage               Enable or disable skills

# Access
/access status               Show this workspace's access mode
/access full                 Full local file and tool access
/access auto                 Auto-approve routine tools
/access ask                  Ask before risky access
/access custom               Use custom config rules

# Context
/context                     Show context and code intelligence state
/context reindex             Rebuild the context index
/compact [instructions]      Compact conversation history now

# Tools
/tools                       Show fixed tool schemas
/tools expand                Expand the latest tool run
/tools compact               Fold long successful tool runs
/tools last                  Show the latest tool trace

# Sessions
/sessions all                Show active and archived sessions
/sessions resume             Attach to a previous session
/sessions new                Start a fresh session

# Daemon
/daemon status               Show daemon and background run state
/daemon queue                Queue a supervised run
/daemon attach               Attach to a supervised run
/daemon detach               Detach a supervised run
/daemon cancel               Cancel a supervised run

# Doctor
/doctor status               Show configuration health
/doctor run                  Probe configured endpoint and optional Omni routes
/doctor tools                Ask the current agent to regress built-in tools

Aliases

Several top-level commands accept friendly aliases so muscle memory from related surfaces still resolves:

Alias	Resolves To
/endpoint, /endpoints	/system