This document is the reference for the visible slash commands available in Coven Code. Commands are invoked by typing /command-name at the REPL prompt.
- Command System Overview
- Session & Navigation
- Model & Provider —
/model,/providers,/connect,/thinking,/effort - Configuration & Settings —
/config,/permissions,/hooks,/mcp - Code & Git —
/commit,/diff,/review,/init,/search - Search & Files —
/link,/attach - Memory & Context —
/memory,/usage,/status - Agents & Tasks —
/familiar,/tasks,/coven goal - Planning & Review —
/plan,ultraplan(CLI) - MCP & Integrations —
/mcp,/skills,/plugin,/chrome - Authentication —
/login,/logout - Display & Terminal —
/incant,/config color - Diagnostics & Info —
/version,/update,/status doctor - Export & Sharing —
/export - Advanced & Internal —
/whisper,/sandbox - Coven Substrate —
/coven,/handoff - Additional Commands — feedback, named CLI commands
- Deprecated Aliases
- Command Availability
Commands are resolved in a priority-ordered registry. When you type a command name, Coven Code checks:
built-in commands -> user command templates -> discovered skills -> plugin commands
Commands support aliases — for example /h, /?, and /help all invoke the same handler.
/command-name [arguments]
Arguments are passed as a single string after the command name.
Aliases: h, ?
Display all available commands with their descriptions. Hidden and setup-only commands are suppressed from the default listing.
/help
/h
/?
Aliases: c, reset, new
Clear the current conversation history and start a fresh session. The session file is retained on disk; only the in-memory message list is cleared.
/clear
Aliases: quit, q
Exit the Coven Code REPL. Equivalent to pressing Ctrl+D. Unsaved session state is flushed before exit.
/exit
/quit
Aliases: r, continue
Resume a previous session from the session store. Displays a list of recent sessions with timestamps and summaries. Select one to restore its message history and file state.
/resume
/resume <session-id>
Aliases: remote
Show or manage conversation sessions. Without arguments, shows the current session status (including the remote session URL when a bridge is active).
/session — show current session status
/session list — list recent sessions
/session rename <new-name> — rename the current session
/session fork [name] — fork into a new independent session
/session branch <name> — create a branch of the conversation at this point
/session tag <tag> — toggle a searchable tag on the current session
/session add-dir <path> — add a directory to the allowed workspace paths
/session rename absorbs the former standalone /rename command. The new name is used in session listings and exports. The former /fork, /branch, /tag, and /add-dir commands fold in here as subcommands (see Deprecated Aliases).
Fork the current session into a new independent session that begins from the current conversation state. Useful for exploring two different approaches without losing either.
/session fork
/session fork <new-session-name>
Single entry point for going back in time. Without arguments, opens an interactive overlay to pick the message to rewind the conversation to. With arguments, rolls back file changes recorded by the shadow-git snapshot system (absorbing the former /undo and /revert).
/rewind — interactive conversation rewind overlay
/rewind list — list assistant turns with recorded file changes
/rewind diff [n] — preview a turn's diff without reverting
/rewind last — revert the most recent assistant turn
/rewind <n> — revert the n-th most recent assistant turn
/rewind <uuid> — revert the turn whose message id starts with <uuid>
/undo and /revert remain hidden compatibility aliases for one release.
Summarize and compress the conversation history to reduce context window usage. The model is asked to produce a dense summary of the prior exchange; that summary replaces the raw messages.
/compact
Open the interactive model picker. Displays a searchable list of available models from all configured providers. The selected model is used for all subsequent inference in the current session.
/model
/model claude-opus-4-5
/model claude-sonnet-4-6
List all configured AI providers and their connection status. Shows provider name, base URL, and whether credentials are present.
/providers
Open the interactive provider picker. The picker offers local Claude Code/ant CLI credential import and Codex CLI login.
/connect
Aliases: think
Configure extended thinking for the current session. Extended thinking allows the model to reason through problems before responding, at the cost of additional tokens.
/thinking
/thinking on
/thinking off
/thinking back [play] [n]
See also /effort for a higher-level interface to thinking depth, and /thinking back below for viewing past thinking traces.
Display the extended-thinking traces from previous model responses in the current session. Only available when extended thinking was used for those responses. Pass a number to view the Nth most-recent trace. Replaces the former /think-back command.
/thinking back — show the most recent thinking trace
/thinking back 2 — show the second most recent thinking trace
/thinking back play — replay the most recent trace as an animated walkthrough
/thinking back play 2 — replay the second most recent trace
/thinking back play absorbs the former /thinkback-play command. Thinking traces appear when the model uses extended thinking mode (see /thinking). If no traces are found, Coven Code suggests enabling extended thinking.
Set the thinking effort level. This is a convenience wrapper over /thinking that maps human-readable levels to token budgets.
| Level | Description |
|---|---|
low |
Minimal thinking; fastest responses |
medium |
Balanced thinking and speed |
high |
Deep reasoning; slower responses |
max |
Maximum token budget for thinking |
/effort low
/effort medium
/effort high
/effort max
See also /effort fast below for switching to the provider's faster model. For a secondary advisor model, see /config advisor.
Toggle fast mode. In fast mode, Coven Code switches to the active provider's smaller, faster model for quick responses. Useful when you want rapid answers and deep reasoning is not required. Replaces the former /fast command.
/effort fast — toggle fast mode on/off
/effort fast on — enable fast mode
/effort fast off — disable fast mode
Setting persists to ~/.coven-code/ui-settings.json.
Aliases: settings
View or modify Coven Code configuration values. Without arguments, renders an interactive settings panel. With arguments, acts as a key-value accessor.
/config
/config get <key>
/config set <key> <value>
/config reset <key>
Common keys:
| Key | Description |
|---|---|
model |
Default model name |
theme |
Color theme name |
vim |
Vim mode enabled (true/false) |
outputStyle |
Output rendering style |
autoApprove |
Auto-approve tool calls |
Subcommands: /config keybindings, /config theme [name], /config output-style [style], /config import, /config advisor [model|off], /config color, /config vim, /config voice, /config statusline, /config terminal-setup — each documented below or under Display & Terminal. Use /splash for the empty-session welcome panel.
/config import imports user-level Claude Code configuration (CLAUDE.md, settings.json) from ~/.claude via an interactive dialog with preview and confirmation. It replaces the former /import-config command.
Open the interactive keybinding configurator. Displays all bound actions with their current shortcuts. Select an action to rebind it. Changes are written to ~/.coven-code/keybindings.json. Replaces the former /keybindings command.
/config keybindings
See keybindings.md for the full keybindings reference.
Show, hide, or toggle the empty-session welcome/splash panel. The setting is written to ~/.coven-code/settings.json under config.showSplash.
/splash — toggle the splash screen
/splash show — show it
/splash hide — hide it
/splash status — show current state
View and manage tool permission rules. Permissions control which tools can run without prompting, which are blocked, and which always require confirmation.
/permissions
/permissions list
/permissions allow <tool-name>
/permissions deny <tool-name>
/permissions reset
Manage event hooks. Hooks are shell commands or scripts that execute when lifecycle events fire (e.g., before/after tool calls, on session start/end).
/hooks
/hooks list
/hooks add <event> <command>
/hooks remove <hook-id>
Available events: pre-tool, post-tool, session-start, session-end, message-send, message-receive.
Inspect Model Context Protocol (MCP) servers and reconnect configured servers. MCP servers expose additional tools and resources to the agent.
/mcp
/mcp list
/mcp status
/mcp auth <name>
/mcp connect <name>
/mcp logs <name>
/mcp resources [name]
/mcp prompts [name]
/mcp get-prompt <name> <prompt> [key=value ...]
Add or remove MCP servers by editing ~/.coven-code/settings.json.
Select how the model's output is rendered in the terminal. Choices include auto, plain, markdown, streaming, and others depending on terminal capabilities. Replaces the former /output-style command.
/config output-style
/config output-style plain
/config output-style markdown
Open the interactive theme picker. Preview and select a color theme for the Coven Code TUI. Replaces the former /theme command.
/config theme
/config theme dark
/config theme light
/config theme solarized
Set or unset a secondary advisor model that provides supplementary suggestions alongside the main model. When set, the advisor model's context is available to improve main-model responses. Replaces the former /advisor command.
/config advisor — show current advisor setting
/config advisor claude-opus-4-6 — set advisor model by name
/config advisor provider/model — set advisor using provider/model format
/config advisor off — disable the advisor
/config advisor unset — disable the advisor
The advisor model persists to ~/.coven-code/settings.json under advisorModel. Model IDs must start with claude- or contain a / (provider/model format).
Configure the status line displayed at the bottom of the TUI. Toggle individual elements such as model name, token count, session name, and git branch.
/config statusline
/config statusline show model
/config statusline hide tokens
Toggle vim keybinding mode on or off. In vim mode the input field behaves like a vim editor (normal/insert/visual modes). Persisted to config.
/config vim
/config vim on
/config vim off
Configure voice input/output. Requires a supported audio backend. Subcommands control microphone selection, TTS voice, and push-to-talk behavior.
/config voice status
/config voice on
/config voice off
Run the terminal capability detection and setup wizard. Checks for true-color support, font ligatures, Unicode rendering, and configures Coven Code accordingly.
/config terminal-setup
Stage and commit changes to the current git repository. The model drafts a commit message based on the diff. You can review and edit the message before confirming.
/commit
/commit "optional message override"
Show file diffs for changes made during the current session. Displays a unified diff of all files Coven Code has written or edited since the session started.
/diff
/diff <file-path>
Review code changes via the model and optionally post the review to the associated GitHub PR. Runs git diff <base>...HEAD (or git diff --cached when no base is given) and sends the diff for a structured review.
/review — review staged changes
/review main — review the diff from main..HEAD
/review origin/main — review against a remote base ref
Variants (the former /security-review and /ultrareview commands fold in here as subcommands):
/review security [path] — security-focused review: vulnerabilities, credential
exposure, injection risks, and other security concerns
/review ultra [path] — exhaustive multi-dimensional review covering security,
performance, maintainability, error handling, test
coverage, API design, and architecture; each finding
is tagged by category and severity
/review comments — fetch comments from the associated GitHub pull
request (replaces the former /pr-comments command)
GitHub posting requires GITHUB_TOKEN (a personal access token with repo scope); the PR number is auto-detected from git remote or supplied via CLAUDE_PR_NUMBER.
Initialize Coven Code project configuration in the current directory. Creates a CLAUDE.md file that acts as persistent project-level context injected at the start of every session.
/init
Search the codebase using natural language or regex patterns. Wraps the GrepTool and GlobTool with a higher-level interface.
/search <query>
/search "TODO" --type ts
/search "function.*export" --regex
Analyze context window usage. Shows a breakdown of tokens consumed by system prompt, conversation history, file contents, and tool results. Helps identify what to compact or drop.
/usage context
Save and manage links in the structured stash. Links are stored in ~/.coven-code/stash.sqlite with title, note, tags, project, and session metadata.
/link <url> [--title ...] [--note ...] [--tags a,b] — save a link
/link add <url> [flags] — same as above
/link list [--tag <tag>] [--project] — list saved links
/link search <term> — search links
/link remove <id> — delete a link
/link list shows each entry's short id, save date, title, URL, and tags. --project limits the listing to links saved from the current repository. The stash is a personal local store and is disabled in hosted review mode.
/link https://docs.rs/tokio --title Tokio docs --tags rust,async
/link list --tag rust
/link remove 1a2b3c4d
Save and manage file attachments in the structured stash. Files are copied into ~/.coven-code/attachments/<id>/, so the stored copy survives even if the original file moves or is deleted. Metadata lives in ~/.coven-code/stash.sqlite alongside saved links.
/attach <path> [--title ...] [--note ...] [--tags a,b] — save a file
/attach add <path> [flags] — same as above
/attach list [--tag <tag>] [--project] — list attachments
/attach search <term> — search attachments
/attach show <id> — show stored/original paths
/attach remove <id> — delete an attachment and its stored copy
Relative paths resolve against the current working directory; ~ expands to the home directory. Paths and flag values containing spaces can be quoted (/attach "My File.pdf" --title "API spec"); unquoted attach paths with spaces also work as long as no flags follow mid-path. Like /link, the stash is disabled in hosted review mode.
/attach ./design/spec.pdf --title API spec --tags design
/attach list --tag design
/attach show 1a2b3c4d
Manage session memory. Memory entries are short notes persisted across sessions. The model can read these at session start to maintain continuity.
/memory
/memory list
/memory add <note>
/memory delete <id>
/memory clear
Display a detailed token usage breakdown for the current session. Shows input tokens, output tokens, cache reads, cache writes, and estimated cost per API call.
/usage
/usage cost
/usage context
/usage stats
In the TUI, /usage stats opens the interactive session statistics dialog (replacing the former /stats command).
Show the total token usage and estimated cost for the current session. Provides a quick summary without the full account/quota context of /usage. In the TUI, /usage stats opens the interactive stats dialog.
/usage cost
For aggregate token / cost / tool statistics across saved sessions, use the stats CLI command: coven-code stats [summary|sessions|tools|daily|session <id>].
Show the current session status. Includes active model, permission mode, thinking config, connected MCP servers, and loaded plugins.
/status
/status doctor — run the diagnostics suite
See /status doctor under Diagnostics & Info for the full diagnostics description.
Aliases: /familiars, /agent
The single surface for familiars and agents. Familiars (from ~/.coven/familiars.toml) and agents (built-in + workspace) resolve through the same runtime agent map and access-tier security under the hood, so one command covers both. Absorbs the former /agent command — which had absorbed /agents and /managed-agents.
/familiar — open the familiars/agents menu (TUI); text overview elsewhere
/familiar status — text overview: current familiar, roster, agents, access tiers
/familiar <name> — switch familiar: persists the choice and immediately applies
the persona's agent definition + tool access tier
/familiar <agent> — show details for a built-in or workspace agent
/familiar info <name> — show details for any familiar or agent
/familiar list — list all visible named agents with access levels
/familiar create [name] — create a new workspace agent
/familiar edit <name> — edit an existing workspace agent
/familiar delete <name> — delete a workspace agent
/familiar reset — clear the active familiar
/familiar reset-roster — erase saved user agents and familiar roster
/familiar managed ... — configure manager-executor agents (see below)
/familiar auto — infer your familiar from the system username
Switching a familiar does three things at once: it changes the TUI mascot, persists the choice to ~/.coven-code/settings.json, and activates the familiar's agent definition — which re-filters the available tools to the familiar's access tier (full, read-only, or search-only). The F2 key opens a quick switcher with the same semantics.
The reset-roster action removes ~/.coven/familiars.toml, custom *.md agent files
from ~/.coven-code/agents/ and the current workspace's .coven-code/agents/,
and clears agents, familiar, and managed_agents settings. It does not
remove built-in agents, plugin packages, sessions, credentials, or history.
To activate an agent at startup, run Coven Code with --agent <name>. See agents.md for defining custom agents and familiars.md for the familiar concept.
Aliases: bashes
Manage tracked background tasks. Tasks are shell commands or model invocations running asynchronously. Monitor progress, fetch output, or stop tasks from this interface.
/tasks
/tasks list
/tasks output <task-id>
/tasks stop <task-id>
Set a durable multi-turn autonomous goal. When a goal is active, Coven Code continues working across turns until the goal is marked complete, paused, or a 200-turn runaway guard fires. Designed for complex, sustained tasks that would otherwise require repeated manual re-prompting. Replaces the former standalone /goal command (which remains a hidden compatibility alias for one release).
/coven goal <objective> — set a new goal and begin working autonomously
/coven goal --tokens 250K <text> — set a goal with a soft token budget cap
/coven goal — show current goal status
/coven goal status — show current goal status
/coven goal pause — pause the active goal
/coven goal resume — resume a paused goal
/coven goal clear — delete the current goal
/coven goal complete — request a completion audit
When the model believes the goal has been achieved, it calls the GoalComplete tool with an audit summary and evidence. Goals can be disabled globally by setting COVEN_CODE_GOALS=0 in your environment.
See Goal System in the advanced guide.
Configure the manager-executor agent architecture, where a manager model delegates subtasks to one or more executor agents working in parallel. Includes budget controls and isolation options. Replaces the former /managed-agents command.
/familiar managed — show current configuration
/familiar managed status — show current configuration
/familiar managed presets — list built-in presets
/familiar managed preset <name> — apply a named preset
/familiar managed setup — show setup instructions
/familiar managed enable — enable managed agents
/familiar managed disable — disable managed agents
/familiar managed reset — remove all managed-agent configuration
/familiar managed configure manager-model <model> — set the manager model
/familiar managed configure executor-model <model> — set the executor model
/familiar managed configure executor-turns <n> — set executor max turns
/familiar managed configure concurrent <n> — set max concurrent executors
/familiar managed configure isolation on|off — toggle executor isolation
/familiar managed configure budget-split shared — shared token pool
/familiar managed configure budget-split percentage:<n> — percentage split (manager gets n%)
/familiar managed configure budget-split fixed:<m>:<e> — fixed USD caps (manager / executor)
/familiar managed budget <amount> — set total budget in USD (0 to clear)
Model format: provider/model (e.g., anthropic/claude-opus-4-6, anthropic/claude-sonnet-4-6). Configuration persists to ~/.coven-code/settings.json under managed_agents.
Preview feature. Behaviour may change across releases.
See Managed Agents in the advanced guide.
Enter plan mode (read-only). In plan mode the model can read files and reason about changes but cannot write, edit, or execute anything. Use this to draft an approach before allowing writes.
/plan
To exit plan mode, use /plan off or the /exit-plan internal action.
Launch the Ultraplan agentic code planner with extended thinking. Like /plan but with an elevated thinking budget to allow more thorough analysis before acting. Ultraplan is a named CLI command — it has no slash form.
coven-code ultraplan [--effort=medium|high|maximum]
For an exhaustive review pass, see /review ultra.
Documented above under Configuration & Settings.
Aliases: skill
List and manage skills. Skills are bundled prompt-commands that extend Coven Code's capabilities without writing code. They appear alongside built-in commands in the registry.
/skills
/skills list
/skills enable <skill-name>
/skills disable <skill-name>
/skills reload
Aliases: plugins
Manage plugins. Plugins are loadable modules that can register new commands, tools, hooks, agents, skills, and MCP server definitions.
/plugin
/plugin list
/plugin info <name>
/plugin enable <name>
/plugin disable <name>
/plugin install <path>
/plugin reload
/plugin reload refreshes the active session plugin registry, hook registry, plugin commands, agents, skills, and in-memory MCP server definitions. New plugin MCP servers are included in the initial MCP connection at startup; if a reload adds a new MCP server after startup, start a new session before expecting its tools in the model tool list.
Browser automation via Chrome DevTools Protocol (CDP). Connects to a running Chrome or Chromium instance and lets Coven Code control it — navigate pages, click elements, fill forms, evaluate JavaScript, and take screenshots.
First, launch Chrome with remote debugging enabled:
chrome --remote-debugging-port=9222 --no-first-runThen:
/chrome connect [--port 9222] — connect to Chrome on the given port (default: 9222)
/chrome navigate <url> — navigate to a URL
/chrome screenshot — take a screenshot, saved to a temp file
/chrome click <selector> — click a CSS selector
/chrome fill <selector> <text> — fill an input field
/chrome eval <js> — evaluate JavaScript and return the result
/chrome disconnect — disconnect from Chrome
Useful for testing web applications, scraping, or automating browser-based workflows without a separate browser-automation tool.
Coven Code supports multiple named accounts per provider — Anthropic (Claude.ai or Console) and Codex (OpenAI ChatGPT subscription). Each login creates a profile under ~/.coven-code/accounts/<provider>/<id>/ and the registry at ~/.coven-code/accounts.json tracks which one is active.
See Authentication Guide for the full story and on-disk layout.
Authenticate with Anthropic or Codex via OAuth PKCE. Opens a browser for the flow and saves tokens under the active profile (or creates a new profile if none exists).
/login — Claude.ai OAuth (Bearer token, default)
/login --console — Console OAuth (creates an API key)
/login --codex — Codex / ChatGPT OAuth
/login --label work — name the new profile "work"
/login --codex --label personal — Codex login, name the profile "personal"
If a profile matching the JWT's email or account_id already exists, that profile is refreshed in place — re-logging-in is idempotent. Use --label to either name a fresh profile or to disambiguate.
Subcommands: /login switch [name] to change the active account and /login refresh to force a token refresh — both documented below.
Remove credentials. By default removes only the active profile for the provider; other stored profiles remain switchable.
/logout — clear active Anthropic profile (drops it from registry)
/logout --codex — clear active Codex profile
/logout --all — purge every Anthropic profile + clear any API key in settings
/logout --codex --all — purge every Codex profile
Switch the active account for a provider. Anthropic by default; pass --codex for Codex. Run /login switch with no arguments to list every stored account and see available profile ids — the active profile in each provider is marked with *. Replaces the former /switch command.
/login switch — list stored accounts across providers
/login switch work — set active Anthropic profile to "work"
/login switch --codex personal — set active Codex profile to "personal"
Sample listing output:
Anthropic:
* personal [pro] kuber@personal.example
work [max] kuber@company.example
Codex:
work kuber@company.example
List every stored account across providers (same listing as /login switch
with no arguments). When several Anthropic profiles point at the same
underlying account — the usual cause is importing the same Claude Code
credential more than once — the listing flags the duplicates.
/accounts — list stored accounts, flagging duplicates
/accounts dedupe — collapse duplicate profiles, keeping the active/freshest
Duplicate profiles all bill the same subscription, so switching between them never escapes a rate limit. The rate-limit recovery dialog offers the same cleanup with a single key when it detects this situation.
Refresh the provider authentication state. Forces a token refresh without full re-authentication. Useful when a session token has expired mid-session. Replaces the former /refresh command.
/login refresh
Theme, output style, statusline, vim mode, and terminal setup are all /config subcommands documented under Configuration & Settings.
Cast a speech incantation — change the model's voice, trading flourish for tokens. The former /caveman, /rocky, and /normal commands fold in here: voices become arguments and /incant off replaces /normal.
/incant <voice> [lite|full|ultra] — cast an incantation at the given intensity
/incant off — lift the active incantation, return to normal speech
Voices:
| Voice | Description |
|---|---|
caveman |
Why use many token when few token do trick. Strips pleasantries, hedging, articles, and transitional phrases (~75% token reduction) |
rocky |
Eridian engineer from Project Hail Mary. Save big token. Good good good. |
Intensity:
| Level | Description |
|---|---|
lite |
Light touch (~40% reduction) |
full |
The default (~75% reduction) |
ultra |
Maximum compression |
/incant caveman — full caveman mode
/incant rocky lite — Rocky grammar, light touch
/incant off — back to normal speech
Set the prompt bar color for the current session. Accepts standard color names or hex values. The color resets when the session ends unless saved via /config.
/config color — show the current prompt color
/config color <name> — set to a named color (e.g., blue, red, green)
/config color #ff6b6b — set to a hex color value
/config color default — reset to the theme default
Run the Coven Code diagnostics suite. Checks configuration integrity, provider connectivity, tool availability, MCP server health, and reports any issues. Replaces the former /doctor command.
/status doctor
Aliases: v
Display the current Coven Code version string and build metadata.
/version
/v
Check for available updates. Queries the GitHub releases API and displays the latest version. If a newer version exists, prints the download URL or upgrade instructions. Does not auto-update. (/upgrade is a hidden compatibility alias.)
/update
Export the current session transcript. Supported formats include Markdown, JSON, and plain text. The output is written to a file or printed to stdout.
/export
/export --format markdown
/export --format json --output session.json
/export --stdout
/export copy [n] — copy a recent response to the clipboard
/export share — upload the session as a shareable gist
The former /copy and /share commands fold in here as subcommands, documented below.
Copy the most recent assistant response to the system clipboard. Pass a number to copy the Nth most-recent response. On Linux a wl-clipboard or xclip backend is used; on macOS and Windows the native clipboard API is used. Replaces the former /copy command.
/export copy — copy the most recent response
/export copy 2 — copy the second most recent response
/export copy N — copy the Nth most recent response
Upload the current session as a secret GitHub gist and return a shareable URL. The session is rendered as a single self-contained HTML file and uploaded via the gh CLI; a viewer URL of the form https://opencoven.github.io/coven-code/session/#<gist-id> is printed. Replaces the former /share command.
/export share
Requires the GitHub CLI (gh) installed and logged in (gh auth login). The viewer base URL can be overridden with COVEN_CODE_SHARE_VIEWER_URL. Secret gists are unlisted but readable by anyone who has the link.
Aliases: btw
Whisper a side question to your familiar without adding it to history. The question goes to the model out-of-band — the response is shown inline but does not become part of the main conversation context. Replaces the former /btw command.
/whisper <question>
/whisper what is the capital of France?
Aliases: sandbox-toggle
Enable or disable sandboxed execution of shell commands. When sandbox mode is on, bash/shell commands run in an isolated environment to limit unintended side effects. Supported on macOS, Linux, and WSL2.
/sandbox — toggle sandbox mode on/off
/sandbox on — enable sandbox mode
/sandbox off — disable sandbox mode
/sandbox status — show current state and excluded patterns
/sandbox exclude <pattern> — add a command pattern to the exclusion list
A restart is recommended after toggling for full effect. On Windows (non-WSL), sandbox mode is not supported.
These commands integrate Coven Code with the local Coven daemon
(~/.coven/coven.sock, contract coven.daemon.v1). They degrade
gracefully when the daemon is absent — the daemon makes coven-code
richer; it is not required.
Drive the local Coven daemon (sessions, harness runs, rituals,
familiars, capability discovery) without leaving the TUI. The
top-level /coven (or /coven status) prints daemon health and the
active session count.
/coven show daemon health
/coven status same as /coven
/coven capabilities daemon capability catalog
/coven familiars list familiar statuses
/coven doctor detect installed harness CLIs
/coven daemon start|status|stop|restart daemon lifecycle
/coven sessions [--all] list active (or all) sessions
/coven info <session-id> full session record
/coven log <session-id> redacted log preview
/coven events <id> [--after N] [--limit M]
paginate session events
/coven events defaults to --limit 50 and prints
End of events stream. when there are no more events. Pass
--limit 0 to fall back to the daemon's default cap.
/coven run <harness> <prompt> launch a new harness session
/coven send <session-id> <text> forward input to a live session
/coven kill <session-id> terminate a live session
/coven attach <session-id> replay/follow a running session
send and kill surface the daemon's structured error codes — when
a session is missing or no longer running you'll see
Session is not running (409 session_not_live) or
Session not found (404 session_not_found) rather than a generic
"daemon offline" message.
/coven summon <session-id> restore an archived session
/coven archive <session-id> archive a non-running session
/coven sacrifice <session-id> permanently delete a session
Sacrifice is irreversible. /coven sacrifice automatically passes
--yes to the underlying coven sacrifice command (mirroring how
the slash invocation acts as the confirmation).
/coven actions <action> [json-args] POST /api/v1/actions
/coven calls [--limit N] read the Coven Calls
delegation ledger
(~/.coven/cave-coven-calls.json)
/coven claim acquire|release|status|heartbeat|canary [args]
parallel-work claim protocol
/coven hooks-install install pre-commit/pre-push
hooks for the claim protocol
/coven adapter list [--json] list configured harness adapters
/coven adapter doctor [id] diagnose configured harness adapters
/coven adapter install <id> install a trusted local adapter recipe
/coven logs prune [--days N] prune session logs
/coven wt <branch> | --list | --doctor | --prune-merged | --prune-stale [DAYS]
worktree management
/coven patch [name] [issue] open the OpenClaw repair flow
/coven pc [status|top|disk|...] macOS system diagnostics
When ~/.coven/coven.sock is missing, every /coven subcommand
returns the same hint:
Coven daemon offline. Try /coven daemon start.
The rest of coven-code keeps working — you just lose substrate
integration.
Non-2xx responses from the daemon are parsed for the
{ "error": { "code", "message" } } envelope and surfaced as
<message> (<status> <code>). See
docs/API-CONTRACT.md in OpenCoven/coven
for the canonical error code list.
Hand off the current session context to a Coven familiar. Sends the
recent transcript to ~/.coven/coven.sock as a new session under
the named familiar, returning the session id.
/handoff <familiar-id>
Requires the Coven daemon to be running and the familiar id to be
defined in ~/.coven/familiars.toml. See
familiars.md for the full familiar concept.
/familiar is documented in full under
Agents & Tasks — it is the unified surface for familiars
and agents. In substrate terms: switching a familiar changes the TUI
mascot live, persists the choice, and activates the familiar's agent
definition (with its access tier) for the session. The roster is read
from ~/.coven/familiars.toml; live familiar status comes from the
daemon when it is running (/coven familiars).
Press F2 at any time to open the familiar switcher popup interactively (the welcome screen's status block hints at this).
The following commands ship with Coven Code but do not have a full section above. They are grouped by purpose.
| Command | Description |
|---|---|
/feedback (alias bug) |
Submit feedback about Coven Code. /feedback report for a bug report. |
/config import |
Import user-level Claude Code configuration (CLAUDE.md, settings.json) from ~/.claude via an interactive dialog with preview and confirmation. |
/plugin reload |
Reload the active session plugin registry, hooks, agents, skills, and MCP definitions. |
These run as coven-code <name> from the shell. Their old slash
adapters (/agents, /add-dir, /branch, /tag, /pr-comments)
are now hidden one-release aliases — in the TUI this functionality
is reached via /session, /familiar, and /review (see
Deprecated Aliases). ultraplan and stats
are CLI-only.
| Command | Description |
|---|---|
agents |
Manage and configure sub-agents. |
agent |
List or inspect named agents. |
add-dir |
Add a directory to the allowed workspace paths. |
branch |
Branch the current conversation. |
tag |
Toggle a searchable session tag. |
pr-comments |
Get comments from a GitHub PR. |
ultraplan |
Launch the Ultraplan agentic code planner with extended thinking. |
stats |
Aggregate token / cost / tool stats across saved sessions (in the TUI, /usage stats opens the stats dialog). |
memory |
Operator controls for hosted/local memory lifecycle: list, expire, redact, delete, and export the tombstone audit ledger. |
The following top-level commands were folded into subcommands. These hidden aliases remain invocable for one release and will be removed afterwards.
| Old command | New form |
|---|---|
/fork |
/session fork [name] |
/branch |
/session branch ... |
/tag |
/session tag ... |
/add-dir |
/session add-dir <path> |
/copy |
/export copy [n] |
/share |
/export share |
/switch |
/login switch [name] |
/refresh |
/login refresh |
/think-back |
/thinking back [play] |
/fast |
/effort fast [on|off] |
/keybindings |
/config keybindings |
/theme |
/config theme [name] |
/output-style |
/config output-style [style] |
/import-config |
/config import |
/advisor |
/config advisor [model|off] |
/agent |
/familiar (alias; one command for familiars and agents) |
/agents |
/familiar [list|create|edit|delete] [name], /familiar reset-roster |
/managed-agents |
/familiar managed ... |
/pr-comments |
/review comments |
/doctor |
/status doctor |
/stats |
/usage stats |
/reload-plugins |
/plugin reload |
/upgrade |
/update |
Not all commands are available in all contexts.
When running with --remote, only a restricted set of commands is available:
session, exit, clear, help, theme, config, usage, plan, keybindings
Over the Remote Control bridge (used by IDE integrations), only local-type commands are forwarded:
compact, clear, usage
Some commands are available only under certain account or platform conditions:
| Command | Restriction |
|---|---|
/effort fast |
Available when a fast-mode model is configured for the active provider |
/sandbox |
Functional on macOS, Linux, WSL2 only; no-op on native Windows |
/config voice |
Requires an audio backend plus WHISPER_ENDPOINT_URL for transcription |
/chrome |
Requires a running Chrome/Chromium instance launched with remote debugging enabled |