| Filename | Latest commit message | Latest commit date |
|---|---|---|
19cc1b12e2
hive-screen-mcp ran `grim` / `wtype` through an unbounded `Command::output()` and spoke RFB to neatvnc with no deadline, so a wedged compositor or a VNC server that accepts and never speaks held the agent's turn forever. Each subprocess now has a 30s limit and is killed when it hits it; each RFB exchange has a 10s limit. Both come back to the agent as the tool's text result, like every other failure in this crate. hive-c0re operator-facing text that described behaviour the code lacks: - `hivectl matrix reset-password` printed a "next: hivectl matrix create-user" hint that fails for every target (agents are refused, a non-agent hits M_USER_IN_USE). The line is gone. - a failed `nixos-container update` appended the container's journal tail, read with `journalctl -M`. Since the job DAG, `update` only runs from the `Swap` node on a stopped container, so the read always came back empty. The helper is removed; the error still points at the build log. - the matrix sweep comment in main.rs said it re-provisions agent token files; `ensure_all` creates no agent accounts or tokens. - the knowledge-pull comments named a webhook caller that no longer exists and claimed a race was "fixed at its source". - `handle_spawn`'s doc and the `mcp_sockets` module doc named callers of `register_agent` / a rollback that do not exist. Refs #4723 |
||
| .. | ||
| src | ||
| Cargo.toml | ||
| README.md | ||
hive-screen-mcp
A stdio MCP server that exposes screenshot, keyboard, and mouse tools to agents
running a Weston Wayland compositor (hyperhive.gui.enable = true).
When to use it
This crate is the bridge for GUI agents — agents that drive graphical
applications (browsers, desktop tools, game UIs) and need to see and interact
with a Weston Wayland display. If your agent does not have hyperhive.gui.enable = true, none of these tools will function.
Tools
| Tool | Mechanism | Requires |
|---|---|---|
screenshot |
grim → PNG on stdout |
grim in PATH, WAYLAND_DISPLAY |
type_text |
wtype (Wayland virtual-keyboard protocol) |
wtype in PATH |
key_press |
wtype -k (XKB keysym syntax) |
wtype in PATH |
mouse_move |
RFB PointerEvent to the neatvnc backend |
HIVE_GUI_VNC_PORT |
mouse_click |
RFB move → button-down → button-up sequence | HIVE_GUI_VNC_PORT |
All tools operate in userspace — no /dev/uinput or kernel bypass.
WAYLAND_DISPLAY and XDG_RUNTIME_DIR are injected by the weston-vnc NixOS
module; HIVE_GUI_VNC_PORT is set by the harness service (defaults to 5900).
Shape
One binary (hive-screen-mcp) launched over stdio. The agent runtime wires it
as an extra MCP server when hyperhive.gui.enable = true. There are no library
exports — it is a self-contained binary crate.
See the crate-root //! docs for the full per-tool implementation notes and
environment variable contract.