+ lHj!a0t$Rt^RIHt^RIHtHt/RRRR.RKORR/bR RRR.RLORR /bR RRRR /bR RRRR/bRRRRR/bRRRRR/bRRRRRR/R^R^RR/bRRRR.RMORR/bRRRRRRR.RNO/RR/bRRRRR/bR RRRR!/bR"RRRRR/R^R^RR#/bR$RRRRR/R^R^RR%/bR&RRR.ROORR'/bR(RRRR)/bR*RRRR+/bR,RRRR-/bR.RRRR//R0RR1RR2/R3RR4RR5/R6RRRR7R8.RR9/R:RR4RR;/R]R?&R@RARRBRCRRDRE]RFR.//t R>]RG&RHRIlt RJ#)PuSchema for the generic `computer_use` tool (model-facing; value is byte-frozen — the schema goes to the model every turn, so prompt-cache parity depends on it). Model-agnostic: any tool-calling model can drive this. Vision-capable models should prefer `capture(mode='som')` then `click(element=N)` — much more reliable than pixel coordinates, which remain supported for models trained on them. ) annotations)AnyDictactiontypestringenum descriptionu Which action to perform. `capture` is free (no side effects). All other actions require approval unless auto-approved. Use `set_value` for select/popup elements and sliders — it selects the matching option directly without opening the native menu (no focus steal).modeuCapture mode. `som` (default) is a screenshot with numbered overlays on every interactable element plus the AX tree — best for vision models, lets you click by element index. `vision` is a plain screenshot. `ax` is the accessibility tree only (no image; useful for text-only models).appaOptional. Limit capture/action to one app (name e.g. 'Safari', or bundle ID). Omitted = frontmost window. app='screen' = composited full-screen grab (image only, no clickable elements); app='desktop' = the OS desktop/shell surface (wallpaper, icons, taskbar) with its elements.pidintegerzqOptional exact process target for action='capture'. Pair with window_id when discovery cannot resolve an X11 app. window_idzOptional exact native window target for action='capture'. Pair with pid when an external cua-driver list_windows lookup has already identified the window.elementzoThe 1-based SOM index returned by the last `capture(mode='som')` call. Strongly preferred over raw coordinates. coordinatearrayitemsminItemsmaxItemszPixel coordinates [x, y] relative to the captured window screenshot (top-left origin). Only use this if no element index is available.buttonzMouse button. Defaults to left. modifiersz%Modifier keys held during the action. from_elementzSource element index (drag). to_elementzTarget element index (drag).from_coordinatez3Source [x,y] (drag; use when no element available). to_coordinatez3Target [x,y] (drag; use when no element available). directionzScroll direction.amountzScroll wheel ticks. Default 3.valuezFor action='set_value': the value to set on the element. For AXPopUpButton / select dropdowns, pass the option's display label (e.g. 'Blue'). For sliders and other AXValue-settable elements, pass the numeric or string value.textz+Text to type (respects the current layout).keyszUKey combo, e.g. 'cmd+s', 'ctrl+alt+t', 'return', 'escape', 'tab'. Use '+' to combine.secondsnumberz wait: seconds to pause (max 30). raise_windowbooleanuOnly for action='focus_app'. If true, brings the window to front (DISRUPTS the user). Default false — input is routed to the app without raising, matching the background co-work model. delivery_mode background foregroundu{For input actions (click, type, key, drag, scroll). `background` (DEFAULT) delivers without raising the window or stealing focus. `foreground` briefly fronts the window then restores focus — a visible change needing its own approval; use it only when a result's verdict tells you to escalate there. Each result's `verdict` carries the next step; follow it rather than guessing.bring_to_frontzOptional and only valid with delivery_mode='foreground'. Explicitly invokes cua-driver's standalone bring_to_front tool before the input; it is never passed as an input property. This persistent focus change has a separate approval scope. Default false. capture_afterzIf true, take a follow-up capture after the action and include it in the response. Saves a round-trip when you need to verify an action's effect.dict[str, Any] _PROPERTIESname computer_useuDrive the desktop via cua-driver — screenshots, mouse, keyboard, scroll, drag — on macOS, Windows, and Linux. Input is background-FIRST, not background-only: the default delivery routes to the target window without stealing the user's cursor or focus (works even on hidden/minimized windows), and when a result's `verdict` says to escalate you climb — pixel coordinates, or delivery_mode='foreground' (briefly fronts the window; separate approval). Each result carries a `verdict` with the next step; follow it — never repeat confirmed input, and re-capture to verify an unverifiable one before retrying. Workflow: action='capture' (mode='som' gives numbered element overlays), then click by `element` index; re-capture after state-changing actions (or pass capture_after=true). Image captures include a shareable `screenshot_path`; deliver it via the platform's MEDIA syntax when the user asks to see it — not for captures used only for control. parametersobject propertiesrequiredCOMPUTER_USE_SCHEMAcV^8dQhRR/#)returnr))formats">/home/ubuntu/.hermes/hermes-agent/tools/computer_use/schema.py __annotate__r8sc\#)z2Return the generic OpenAI function-calling schema.)r1r5r9r7get_computer_use_schemar;s r9N)captureclick double_click right_click middle_clickdragscrollrkey set_valuewait list_apps list_windows focus_app)somvisionax)leftrightmiddle) cmdshiftoptionaltctrlfnwinwindowssupermeta)updownrLrM) __conditional_annotations____doc__ __future__rtypingrrr*__annotations__r1r;)r[s@r7r`sE# i   'i4 ' 7 5iH  * IiZ   3 [ih  Yiiv  .wiD&)$AA G EiX +8Yib H    >#ciHVY 7UVIiJ69m5STKiL&)$AAL MiZ&)$AAL [ih&(F,K]\opiij vy-1QRkil  Kmi| VX}.[ \}i~  c  -1ST  |, 9     MEi ^iX N Z68\; XJW'^$r9