Enum Command 

pub enum Command {
    SessionOpen {
        policy: Option<String>,
    },
    SessionClose,
    Open {
        url: String,
    },
    Close {
        page: String,
    },
    View {
        page: String,
        full: bool,
    },
    Read {
        page: String,
        r: u32,
    },
    Act {
        page: String,
        r: u32,
        op: String,
        value: Option<String>,
        token: String,
        group: Option<String>,
    },
    Find {
        query: String,
    },
    Wait {
        page: String,
        cond: String,
        value: Option<String>,
        timeout: u64,
    },
    Status,
    Extract {
        page: String,
        schema: String,
        token: String,
    },
    Mark {
        page: String,
        r: u32,
        name: String,
        token: String,
    },
    Annotate {
        target: String,
        key: String,
        value: Option<String>,
    },
    Log {
        page: Option<String>,
        group: Option<String>,
        since: Option<i64>,
        limit: Option<i64>,
    },
    Skill {
        sub: Option<String>,
        name: Option<String>,
    },
    Capture {
        page: String,
        r: Option<u32>,
        full_page: bool,
        base64: bool,
    },
    Viewport {
        page: String,
        spec: String,
        dpr: u32,
    },
    Layout {
        page: String,
        refs: Vec<u32>,
    },
    Auth {
        sub: String,
        rest: Vec<String>,
    },
    Inspect {
        page: String,
        kind: String,
        rest: Vec<String>,
        since: Option<String>,
        level: Option<String>,
        status: Option<String>,
        max: Option<String>,
        full: bool,
        unsafe_log: bool,
    },
    MoveTo {
        page: String,
        x: f64,
        y: f64,
        mode: String,
    },
    ClickAt {
        page: String,
        x: f64,
        y: f64,
        token: String,
        mode: String,
    },
    HoverAt {
        page: String,
        x: f64,
        y: f64,
        mode: String,
    },
    Drag {
        page: String,
        x1: f64,
        y1: f64,
        x2: f64,
        y2: f64,
        token: String,
        mode: String,
    },
    PromptInput {
        page: String,
        r: u32,
        message: String,
        secret: bool,
        token: String,
        group: Option<String>,
    },
    PromptConfirm {
        page: String,
        message: String,
    },
    PromptInputQueue {
        page: String,
        r: u32,
        message: String,
        secret: bool,
        token: String,
        group: Option<String>,
        timeout_ms: u64,
    },
    Pending {
        sub: PendingSub,
    },
    Serve {
        stop: bool,
    },
    Mcp,
}

The subcommand vocabulary. One variant per primitive.

Each variant has a visible_alias short form for token economy in agent contexts (mirrors agented’s s / i / d / w / br pattern). Long forms remain for human readers.

Variants

SessionOpen

1. Open a new session.

Fields

policy: Option<String>

SessionClose

2. Close the active session.

Open

3. Open a page navigated to URL.

Fields

url: String

Close

4. Close a page.

Fields

page: String

View

5. View a page (delta by default; --full re-baselines).

Fields

page: String

full: bool

Read

6. Read the full text of a ref.

Fields

page: String

r: u32

Act

7. Perform an action on a ref.

Fields

page: String

r: u32

op: String

value: Option<String>

token: String

group: Option<String>

Find

8. Search across pages in the session.

Fields

query: String

Wait

9. Wait for a condition on a page.

Fields

page: String

cond: String

value: Option<String>

timeout: u64

Status

13. Status summary for the active session (or all sessions).

Extract

10. Extract structured data using a known schema.

Fields

page: String

schema: String

token: String

Mark

11. Persist a ref as a named anchor.

Fields

page: String

r: u32

name: String

token: String

Annotate

12. Attach a (key, value) annotation to a target (one of ref:N, mark:NAME, or page).

Fields

target: String

key: String

value: Option<String>

Log

14. Slice the audit log.

Fields

page: Option<String>

group: Option<String>

since: Option<i64>

limit: Option<i64>

Skill

15. Skill management. Subcommand: list or show <name>. (M6 adds <name> [args] execution.)

Fields

sub: Option<String>

name: Option<String>

Capture

16. Capture a screenshot. Defaults to viewport scope; pass a ref to capture an element, or --full-page.

Fields

page: String

r: Option<u32>

full_page: bool

base64: bool

Emit the PNG bytes as base64 on the response body (instead of just a path on disk). Lets MCP-driven agents see the screenshot inline; the on-disk PNG is still written.

Viewport

17. Set the viewport. spec is a preset (e.g. mobile, desktop) or WxH (e.g. 1280x720).

Fields

page: String

spec: String

dpr: u32

Layout

18. Compute layout boxes for one or more refs.

Fields

page: String

refs: Vec<u32>

Auth

19. Auth blob management. Subcommand: save <page> <name>, load <page> <name>, list, or clear <name>.

Fields

sub: String

rest: Vec<String>

Inspect

20. Inspect engine state — console, network, request detail, storage, scripts, dom, performance. The first positional is the page id; the second is the kind. Trailing positionals are kind-specific (e.g. request <seq>, eval <expr>, storage <scope>, script <seq>).

Fields

page: String

kind: String

rest: Vec<String>

since: Option<String>

level: Option<String>

status: Option<String>

max: Option<String>

full: bool

unsafe_log: bool

MoveTo

Move the cursor to (x, y) along a humanized Bezier path. Native trusted-event dispatch on macOS; ENGINE_UNSUPPORTED elsewhere until M7 wires GDK / CDP input.

Fields

page: String

x: f64

y: f64

mode: String

ClickAt

Click at (x, y). Trusted on macOS (isTrusted = true).

Fields

page: String

x: f64

y: f64

token: String

mode: String

HoverAt

Hover at (x, y).

Fields

page: String

x: f64

y: f64

mode: String

Drag

Drag from (x1, y1) to (x2, y2).

Fields

page: String

x1: f64

y1: f64

x2: f64

y2: f64

token: String

mode: String

PromptInput

Prompt the user (in the terminal that ran vs) for a value, then fill it into a ref. The value is read from tty by the CLI and shipped to the daemon over the local socket; the agent that invoked vs prompt-input never sees the bytes. Use --secret for passwords, TANs, and other credentials (terminal echo off).

Fields

page: String

r: u32

message: String

secret: bool

token: String

group: Option<String>

PromptConfirm

Block until the user presses Enter in the terminal. Returns ok on confirm or ! ABORTED if the user sends EOF / Ctrl-C. Use as a human-in-loop gate before a sensitive vs act click.

Fields

page: String

message: String

PromptInputQueue

Wire-only vs_prompt_input variant — does NOT read from the local tty. Used by vs mcp so an MCP-driven agent can enqueue a prompt the local user fulfills via vs pending fulfill. Hidden from --help: only the MCP server wires it.

Fields

page: String

r: u32

message: String

secret: bool

token: String

group: Option<String>

timeout_ms: u64

Timeout in milliseconds before the daemon gives up waiting for vs pending fulfill <id> (default 5 min).

Pending

List / fulfill / cancel pending vs_prompt_input entries queued by an MCP-driven agent. Use vs pending list to see what’s waiting, vs pending fulfill <id> to type the value at the local tty (--secret hides echo), vs pending cancel <id> to abort.

Fields

sub: PendingSub

Serve

Run the daemon in this process. The vs binary doubles as the daemon — vs serve is what auto-spawn re-execs when the socket is missing. SIGINT shuts down cleanly.

Fields

stop: bool

Send SIGTERM to the running daemon (PID file at ~/.vibesurfer/daemon.pid) and wait for the socket to disappear. Returns immediately if no daemon is running.

Mcp

Run the MCP (Model Context Protocol) server over stdio. Speaks JSON-RPC 2.0; each of the 19 vibesurfer primitives is exposed as one MCP tool. Wire to Claude Desktop / Claude Code by configuring vs mcp as the server command.

Implementations

impl Command

pub fn to_request(&self, session_id: Option<&str>) -> Result<Request>

Build the wire Request for this subcommand. Returns None for commands that the CLI handles locally (none yet).

pub fn needs_session(&self) -> bool

True if this subcommand requires an active session.

Trait Implementations

impl Debug for Command

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl FromArgMatches for Command

fn from_arg_matches(__clap_arg_matches: &ArgMatches) -> Result<Self, Error>

Instantiate Self from ArgMatches, parsing the arguments as needed.

fn from_arg_matches_mut( __clap_arg_matches: &mut ArgMatches, ) -> Result<Self, Error>

Instantiate Self from ArgMatches, parsing the arguments as needed.

fn update_from_arg_matches( &mut self, __clap_arg_matches: &ArgMatches, ) -> Result<(), Error>

Assign values from ArgMatches to self.

fn update_from_arg_matches_mut<'b>( &mut self, __clap_arg_matches: &mut ArgMatches, ) -> Result<(), Error>

Assign values from ArgMatches to self.

impl Subcommand for Command

fn augment_subcommands<'b>(__clap_app: Command) -> Command

Append to Command so it can instantiate Self via FromArgMatches::from_arg_matches_mut

fn augment_subcommands_for_update<'b>(__clap_app: Command) -> Command

Append to Command so it can instantiate self via FromArgMatches::update_from_arg_matches_mut

fn has_subcommand(__clap_name: &str) -> bool

Test whether Self can parse a specific subcommand