Agent Zero AI framework
Split macOS computer-use backend guidance
Alessandro Fraue7cb3aa3fa42
diff --git a/plugins/_a0_connector/prompts/agent.system.tool.computer_use_remote.md b/plugins/_a0_connector/prompts/agent.system.tool.computer_use_remote.mdindex e5d7bb2bd..3b4ed31d7 100644--- a/plugins/_a0_connector/prompts/agent.system.tool.computer_use_remote.md+++ b/plugins/_a0_connector/prompts/agent.system.tool.computer_use_remote.md@@ -8,7 +8,9 @@ This is the only desktop-control path for the user's connected host/local computIf the tool reports no CLI, disabled computer use, or `COMPUTER_USE_REARM_REQUIRED`, stop and tell the user to run `/computer-use on` in A0 CLI and approve any host permission prompt.-Call `start_session` before screen-driven tasks. Use `status` for state only, `capture` for screenshots without an action, and `stop_session` when the desktop task is complete. Interactive actions should use normalized global-screen coordinates from the most recent capture.+Call `start_session` before screen-driven tasks. Use `status` for state only, `capture` for screenshots without an action, and `stop_session` when the desktop task is complete. Interactive coordinate actions should use normalized global-screen coordinates from the most recent capture.++Some actions are backend-specific. Use `ax_snapshot` and `ax_action` only when `status` or `start_session` reports a macOS backend with Accessibility-tree features such as `accessibility-tree-snapshot` or `accessibility-structural-targeting`. For macOS structural targeting details, load and follow skill `host-computer-use-macos`; do not apply AX guidance to non-macOS backends.State-changing actions automatically attach a fresh screen after they run. Treat key presses, clicks, scrolling, typing, and window-manager shortcuts as attempts, not success: inspect the latest attached screen, or one explicit `capture` if it is unclear or unchanged, before saying the requested outcome happened. If the tool says a screen was attached but you cannot actually inspect the image, stop and report that visual verification is unavailable; do not continue by assuming the host state. For Ubuntu/GNOME/Wayland hide-window tasks, prefer `Super+H` (`{"action":"key","keys":["Super","H"]}`) for the active window; do not use `Alt+F9` as the primary hide/minimize shortcut because it often leaves the window visible. A `type` result only proves keystrokes were sent; it does not prove the window was hidden or that text landed in the intended place.@@ -22,13 +24,18 @@ State-changing actions automatically attach a fresh screen after they run. Treat```Required argument:-- `action`: one of `start_session`, `status`, `capture`, `move`, `click`, `scroll`, `key`, `type`, `stop_session`+- `action`: one of `start_session`, `status`, `capture`, `ax_snapshot`, `ax_action`, `move`, `click`, `scroll`, `key`, `type`, `stop_session`Optional arguments by action:- `session_id`: session returned by `start_session`- `x`, `y`: normalized `[0,1]` global-screen coordinates for `move` and `click`- `button`: `left`, `right`, or `middle` for `click`- `count`: click count for `click`+- `max_depth`, `max_nodes`: optional bounds for backend-gated `ax_snapshot`+- `path`: element path from `ax_snapshot` for backend-gated `ax_action`+- `target`: semantic element target for backend-gated `ax_action`, for example role/title/description/value/identifier+- `operation` or `ax_action`: backend-gated AX operation such as `press`, `focus`, or `set_value`+- `value` or `text`: value for backend-gated `ax_action` with `set_value`- `dx`, `dy`: scroll amounts for `scroll`- `key` or `keys`: key press value for `key`- `text`: text to type for `type`diff --git a/plugins/_a0_connector/skills/host-computer-use-macos/SKILL.md b/plugins/_a0_connector/skills/host-computer-use-macos/SKILL.mdnew file mode 100644index 000000000..fdd04a835--- /dev/null+++ b/plugins/_a0_connector/skills/host-computer-use-macos/SKILL.md@@ -0,0 +1,83 @@+---+name: host-computer-use-macos+description: Backend-specific macOS guidance for `computer_use_remote`. Load after `status` or `start_session` reports backend_family/backend_id `macos` or Accessibility-tree features. Covers AX structural targeting, macOS window actions, and screenshot verification.+---++# Host Computer Use - macOS++Use this after `host-computer-use` when the connected A0 CLI reports the macOS computer-use backend.++Do not use this skill for Linux, Windows, Xpra, Docker, or browser-only tasks. If the backend is not macOS or does not advertise Accessibility-tree support, skip AX actions and follow the generic host computer-use rules.++## macOS AX Targeting++macOS backends can advertise structural Accessibility features:++- `accessibility-tree-snapshot`+- `accessibility-structural-targeting`+- `accessibility-element-click`++When these features are present, prefer structural targeting over pixel clicks for named controls such as buttons, menu items, text fields, sheets, alerts, toolbar items, and sidebar rows.++Use `ax_snapshot` to inspect the frontmost app's bounded Accessibility tree:++```json+{+ "tool_name": "computer_use_remote",+ "tool_args": {+ "action": "ax_snapshot",+ "max_depth": 4,+ "max_nodes": 200+ }+}+```++The snapshot returns element paths, roles, labels, frames, enabled/focused state, actions, and child nodes. Use it to choose an element, not as final visual proof.++Use `ax_action` for a structural action:++```json+{+ "tool_name": "computer_use_remote",+ "tool_args": {+ "action": "ax_action",+ "target": {+ "role": "AXButton",+ "title": "OK"+ },+ "operation": "press"+ }+}+```++Supported operations are:++- `press`: activate a button, menu item, checkbox, or similar control+- `focus`: focus a text field or focusable element+- `set_value`: set a value; pass `value` or `text`++Targeting options:++- Prefer a semantic `target` when the control has a stable title, description, value, identifier, role, or subrole.+- Use a `path` returned by the latest `ax_snapshot` only while the UI is unchanged.+- If an action reports ambiguity, take a fresh snapshot and narrow the target with role plus title/description/identifier.+- If an action reports a missing target, take a fresh snapshot before trying coordinates.++AX actions are attempts, not proof. They attach a fresh screenshot after state-changing actions; inspect that image before saying the requested outcome happened.++## macOS Window Actions++For active-app window tasks, macOS shortcuts are usually:++- `Command+H`: hide the active app+- `Command+M`: minimize the active window++Treat these as attempts. After any hide or minimize shortcut, inspect the fresh screenshot. If the target window is still visible, say the attempt failed or switch strategy; do not type follow-up text into the focused app unless the screenshot clearly shows the intended target.++When a visible close/minimize/full-screen button or menu item is accessible in the AX tree, prefer `ax_snapshot` plus `ax_action` over a small coordinate click.++## Permissions++macOS Screen Recording affects screenshots. Accessibility/Input Monitoring affect structural targeting and input. If `computer_use_remote` returns `COMPUTER_USE_REARM_REQUIRED`, `COMPUTER_USE_APPROVAL_REQUIRED`, or `status=rearm required`, stop immediately and ask the user to run `/computer-use on` in the A0 CLI and approve the macOS prompt if shown.++Do not bypass a permission failure with server screenshots, Docker commands, `linux-desktop`, or browser fallbacks.diff --git a/plugins/_a0_connector/skills/host-computer-use/SKILL.md b/plugins/_a0_connector/skills/host-computer-use/SKILL.mdindex 20b1e7eaf..ccb501ce3 100644--- a/plugins/_a0_connector/skills/host-computer-use/SKILL.md+++ b/plugins/_a0_connector/skills/host-computer-use/SKILL.md@@ -57,8 +57,9 @@ Use:Arguments:-- `action`: `start_session`, `status`, `capture`, `move`, `click`, `scroll`, `key`, `type`, `stop_session`+- `action`: `start_session`, `status`, `capture`, `ax_snapshot`, `ax_action`, `move`, `click`, `scroll`, `key`, `type`, `stop_session`- `session_id`: optional after `start_session`+- `ax_snapshot`, `ax_action`: backend-gated structural accessibility actions; use only when backend metadata advertises matching support, and load the backend-specific skill first- `move`: `x`, `y` normalized to `[0,1]`- `click`: optional `x`, `y`, optional `button` (`left`, `right`, `middle`), optional `count`- `scroll`: `dx`, `dy`@@ -72,16 +73,22 @@ If any tool result contains `COMPUTER_USE_REARM_REQUIRED` or `status=rearm requi## Core Loop1. Call `start_session` first.-2. Decide from the latest screenshot, not from memory.-3. Interactive actions (`move`, `click`, `scroll`, `key`, `type`) already attach a fresh screenshot after they run; inspect it before claiming the requested outcome succeeded.-4. Use `status` for state without starting a session.-5. Use `capture` only when you need another screenshot without taking an action.+2. Read the returned `backend_id`, `backend_family`, and `features`; load a backend-specific Computer Use skill when the task needs backend-only affordances.+3. Decide final success from the latest screenshot, not from memory.+4. Interactive actions (`ax_action`, `move`, `click`, `scroll`, `key`, `type`) already attach a fresh screenshot after they run; inspect it before claiming the requested outcome succeeded.+5. Use `status` for state without starting a session.+6. Use `capture` only when you need another screenshot without taking an action.++## Backend Skills++- If the backend is macOS or features include `accessibility-tree-snapshot` / `accessibility-structural-targeting`, load `host-computer-use-macos` before using `ax_snapshot` or `ax_action`.+- Do not use backend-specific actions just because their argument names exist in the generic contract. Treat them as unavailable unless the connected CLI advertises the matching feature.## Operating Rules- Only the latest screenshot or a definitive tool result counts as evidence.- If a tool result says a screenshot was attached but you cannot actually see the image, stop and report that visual verification is unavailable. Do not continue with another action from an assumed host state.-- The current API uses normalized global screen coordinates; do not assume window ids, element indexes, background-safe input, or semantic click targets unless the runtime explicitly advertises them.+- Outside advertised AX support, use normalized global screen coordinates; do not assume window ids, element indexes, background-safe input, or semantic click targets unless the runtime explicitly advertises them.- Prefer accessibility and semantic UI paths first: shortcuts, command palettes, menu accelerators, address/search bars, focus traversal, and other keyboard-accessible controls.- Prefer `key` and `type` over pointer actions whenever a reliable keyboard path exists.- When a menu or popup is open, treat it as the active UI and prefer keyboard navigation over clicking small transient rows by coordinate.@@ -101,7 +108,7 @@ If any tool result contains `COMPUTER_USE_REARM_REQUIRED` or `status=rearm requi- Try keyboard scrolling first: `page_down`, `page_up`, `space`, `shift+space`, arrows, `home`, or `end`.- Use `scroll` when the desired pane is already active or keyboard scrolling cannot target it.-- Treat `move` and `click` as last-resort actions for controls that cannot be reached through keyboard, accessibility, browser, or app-native tooling.+- Treat `move` and `click` as last-resort actions for controls that cannot be reached through backend-specific structural targeting, keyboard, browser, or app-native tooling.- Before clicking, make sure the latest screenshot makes the target unambiguous. Use one deliberate click, then reassess from the fresh screenshot.## Control Signalsdiff --git a/plugins/_a0_connector/tools/computer_use_remote.py b/plugins/_a0_connector/tools/computer_use_remote.pyindex f7edcee2b..0f7e3dc25 100644--- a/plugins/_a0_connector/tools/computer_use_remote.py+++ b/plugins/_a0_connector/tools/computer_use_remote.py@@ -34,6 +34,7 @@ REARM_REQUIRED_DEFAULT_MESSAGE = ()_AUTO_CAPTURE_ACTIONS = {"start_session",+ "ax_action","move","click","scroll",@@ -46,6 +47,7 @@ _SETTLE_DELAY_CLICK = 0.35_SETTLE_DELAY_SCROLL = 0.35_SETTLE_DELAY_KEY = 0.2_SETTLE_DELAY_TYPE = 0.25+_SETTLE_DELAY_AX_ACTION = 0.25_SETTLE_DELAY_GLOBAL_FOCUS = 0.45_SETTLE_DELAY_PLAIN_ENTER = 0.3_SETTLE_DELAY_SUBMIT = 0.45@@ -54,6 +56,8 @@ _SUPPORTED_ACTIONS = {"start_session","status","capture",+ "ax_snapshot",+ "ax_action","move","click","scroll",@@ -72,7 +76,8 @@ class ComputerUseRemote(Tool):return Response(message=("action is required and must be one of: "- "start_session, status, capture, move, click, scroll, key, type, stop_session"+ "start_session, status, capture, ax_snapshot, ax_action, "+ "move, click, scroll, key, type, stop_session"),break_loop=False,)@@ -255,6 +260,8 @@ class ComputerUseRemote(Tool):return _SETTLE_DELAY_CLICKif action == "scroll":return _SETTLE_DELAY_SCROLL+ if action == "ax_action":+ return _SETTLE_DELAY_AX_ACTIONif action == "type" and self._coerce_bool(self.args.get("submit")):return _SETTLE_DELAY_SUBMITif action == "type":@@ -310,6 +317,24 @@ class ComputerUseRemote(Tool):payload["text"] = self.args.get("text", "")if self._coerce_bool(self.args.get("submit")):payload["submit"] = True+ elif action == "ax_snapshot":+ if "max_depth" in self.args:+ payload["max_depth"] = self._coerce_int(self.args.get("max_depth"), name="max_depth")+ if "max_nodes" in self.args:+ payload["max_nodes"] = self._coerce_int(self.args.get("max_nodes"), name="max_nodes")+ elif action == "ax_action":+ target = self.args.get("target")+ if isinstance(target, dict):+ payload["target"] = dict(target)+ if "path" in self.args:+ payload["path"] = self.args.get("path")+ operation = self.args.get("operation", self.args.get("ax_action", self.args.get("name")))+ if operation is not None:+ payload["operation"] = operation+ if "value" in self.args:+ payload["value"] = self.args.get("value")+ if "text" in self.args:+ payload["text"] = self.args.get("text", "")return payload@@ -329,13 +354,27 @@ class ComputerUseRemote(Tool):if action == "capture":summary = self._record_capture(data)return f"Current screen attached: {summary} {CAPTURE_VERIFICATION_NOTE}"+ if action == "ax_snapshot":+ return self._format_ax_snapshot(data)+ if action == "ax_action":+ target = data.get("target") if isinstance(data.get("target"), dict) else {}+ operation = str(data.get("operation") or "?")+ path = target.get("path", "?")+ return f"Performed AX {operation} on {self._ax_target_label(target)} path={path}."if action == "status":return self._format_status(data)if action == "start_session":- return (+ message = (f"Computer-use session started: session_id={data.get('session_id', '?')} "f"size={data.get('width', '?')}x{data.get('height', '?')}")+ backend_details = self._format_backend_details(data)+ if backend_details:+ message = f"{message}, {backend_details}"+ skill_hint = self._backend_skill_hint(data)+ if skill_hint:+ return f"{message}.{skill_hint}"+ return messageif action == "stop_session":return "Computer-use session stopped."if action == "move":@@ -373,14 +412,54 @@ class ComputerUseRemote(Tool):return f"{code}: {error}"return error+ def _format_backend_details(self, data: dict[str, Any]) -> str:+ backend_id = str(data.get("backend_id", "") or "").strip()+ backend_family = str(data.get("backend_family", "") or "").strip()+ features = self._backend_features(data)+ parts: list[str] = []+ if backend_id:+ backend_text = backend_id+ if backend_family:+ backend_text = f"{backend_text}/{backend_family}"+ parts.append(f"backend={backend_text}")+ if features:+ parts.append(f"features={', '.join(features)}")+ return ", ".join(parts)++ def _backend_features(self, data: dict[str, Any]) -> list[str]:+ raw_features = data.get("features") or []+ if not isinstance(raw_features, (list, tuple, set)):+ return []+ features: list[str] = []+ for feature in raw_features:+ text = str(feature or "").strip()+ if text:+ features.append(text)+ return features++ def _backend_skill_hint(self, data: dict[str, Any]) -> str:+ backend_id = str(data.get("backend_id", "") or "").strip().lower()+ backend_family = str(data.get("backend_family", "") or "").strip().lower()+ features = {feature.lower() for feature in self._backend_features(data)}+ has_macos_ax = bool(+ features+ & {+ "accessibility-tree-snapshot",+ "accessibility-structural-targeting",+ }+ )+ if backend_id == "macos" or backend_family == "macos" or has_macos_ax:+ return (+ " Load skill `host-computer-use-macos` before using macOS AX "+ "structural actions."+ )+ return ""+def _format_status(self, data: dict[str, Any]) -> str:status = str(data.get("status", "unknown") or "unknown")trust_mode = str(data.get("trust_mode", "") or "")- backend_id = str(data.get("backend_id", "") or "").strip()- backend_family = str(data.get("backend_family", "") or "").strip()active_contexts = data.get("active_contexts") or []active_text = ", ".join(str(item) for item in active_contexts) if active_contexts else "none"- backend_text = ""rearm_guidance = ""if status == "rearm required":detail = str(data.get("last_error") or "").strip()@@ -394,20 +473,37 @@ class ComputerUseRemote(Tool):"is not armed. ""Stop using computer_use_remote until the user re-arms it.")- if backend_id:- backend_text = backend_id- if backend_family:- backend_text = f"{backend_text}/{backend_family}"- if backend_text:+ backend_details = self._format_backend_details(data)+ if backend_details:return (f"Computer use status={status}, trust_mode={trust_mode or 'unknown'}, "- f"backend={backend_text}, active_contexts={active_text}.{rearm_guidance}"+ f"{backend_details}, active_contexts={active_text}."+ f"{self._backend_skill_hint(data)}{rearm_guidance}")return (f"Computer use status={status}, trust_mode={trust_mode or 'unknown'}, "f"active_contexts={active_text}.{rearm_guidance}")+ def _format_ax_snapshot(self, data: dict[str, Any]) -> str:+ app = data.get("app") if isinstance(data.get("app"), dict) else {}+ tree = data.get("tree") if isinstance(data.get("tree"), dict) else {}+ app_name = str(app.get("name") or app.get("bundle_id") or "frontmost app")+ node_count = data.get("node_count", "?")+ truncated = " truncated" if data.get("truncated") else ""+ root_label = self._ax_target_label(tree)+ return (+ f"AX snapshot for {app_name}: {node_count} node(s){truncated}. "+ f"Root {root_label}. Use path or semantic target fields with ax_action."+ )++ def _ax_target_label(self, target: dict[str, Any]) -> str:+ role = str(target.get("role") or "element")+ title = str(target.get("title") or target.get("description") or target.get("identifier") or "").strip()+ if title:+ return f"{role} {title!r}"+ return role+def _record_capture(self, data: dict[str, Any]) -> str:display_ref, resolved_capture_id = self._resolve_capture_ref(data)width = data.get("width", "?")diff --git a/tests/test_a0_connector_prompt_gating.py b/tests/test_a0_connector_prompt_gating.pyindex 3e7c54476..044b07d69 100644--- a/tests/test_a0_connector_prompt_gating.py+++ b/tests/test_a0_connector_prompt_gating.py@@ -345,6 +345,14 @@ def test_remote_affordance_skills_parse():/ "host-computer-use"/ "SKILL.md")+ macos_computer_skill = _parse_skill_frontmatter(+ PROJECT_ROOT+ / "plugins"+ / "_a0_connector"+ / "skills"+ / "host-computer-use-macos"+ / "SKILL.md"+ )assert not legacy_connector_skill.exists()assert text_editor_skill["name"] == "host-file-editing"@@ -358,6 +366,8 @@ def test_remote_affordance_skills_parse():assert "Use instead of linux-desktop" in computer_skill["description"]assert "host computer" in computer_skill["triggers"]assert "Ubuntu Wayland desktop" in computer_skill["triggers"]+ assert macos_computer_skill["name"] == "host-computer-use-macos"+ assert "macOS guidance" in macos_computer_skill["description"]def test_remote_tool_stubs_are_self_contained_and_reference_per_tool_skills():@@ -372,6 +382,14 @@ def test_remote_tool_stubs_are_self_contained_and_reference_per_tool_skills():/ "host-computer-use"/ "SKILL.md").read_text(encoding="utf-8")+ macos_computer_skill = (+ PROJECT_ROOT+ / "plugins"+ / "_a0_connector"+ / "skills"+ / "host-computer-use-macos"+ / "SKILL.md"+ ).read_text(encoding="utf-8")assert "optionally load skill `host-file-editing`" in text_stubassert "optionally load skill `host-code-execution`" in exec_stub@@ -379,8 +397,12 @@ def test_remote_tool_stubs_are_self_contained_and_reference_per_tool_skills():assert '"tool_name": "code_execution_remote"' in exec_stubassert '"tool_name": "computer_use_remote"' in computer_stubassert "load and follow skill `host-computer-use`" in computer_stub+ assert "host-computer-use-macos" in computer_stubassert "Do not substitute the `linux-desktop` skill" in computer_stubassert '"tool_name": "computer_use_remote"' in computer_skill+ assert '"tool_name": "computer_use_remote"' in macos_computer_skill+ assert "ax_snapshot" in macos_computer_skill+ assert "ax_snapshot`/`ax_action` are structural Accessibility targeting" not in computer_skillassert "Availability, backend support, and trust mode are checked when the tool runs" in computer_stubassert "not `code_execution_tool`" in exec_stubassert "not to" in exec_stubdiff --git a/tests/test_skills_runtime.py b/tests/test_skills_runtime.pyindex 1f3aea3ab..f37a2aa46 100644--- a/tests/test_skills_runtime.py+++ b/tests/test_skills_runtime.py@@ -291,6 +291,7 @@ def test_renamed_skills_use_standard_frontmatter_only():PROJECT_ROOT / "skills" / "scheduled-tasks" / "SKILL.md",PROJECT_ROOT / "plugins" / "_a0_connector" / "skills" / "host-code-execution" / "SKILL.md",PROJECT_ROOT / "plugins" / "_a0_connector" / "skills" / "host-computer-use" / "SKILL.md",+ PROJECT_ROOT / "plugins" / "_a0_connector" / "skills" / "host-computer-use-macos" / "SKILL.md",PROJECT_ROOT / "plugins" / "_a0_connector" / "skills" / "host-file-editing" / "SKILL.md",PROJECT_ROOT / "plugins" / "_a0_connector" / "skills" / "setup-a0-cli" / "SKILL.md",PROJECT_ROOT / "plugins" / "_browser" / "skills" / "browser-automation" / "SKILL.md",diff --git a/tests/test_tool_action_contracts.py b/tests/test_tool_action_contracts.pyindex 8ded5a19b..e3289d56b 100644--- a/tests/test_tool_action_contracts.py+++ b/tests/test_tool_action_contracts.py@@ -86,6 +86,50 @@ def _load_skills_tool(monkeypatch, skill_root: Path):return importlib.import_module("tools.skills_tool")+def _load_computer_use_remote_tool(monkeypatch):+ _install_tool_stub(monkeypatch)++ history_stub = types.ModuleType("helpers.history")++ class _RawMessage(dict):+ def __init__(self, raw_content, preview):+ super().__init__(raw_content=raw_content, preview=preview)++ history_stub.RawMessage = _RawMessage+ monkeypatch.setitem(sys.modules, "helpers.history", history_stub)++ print_style_stub = types.ModuleType("helpers.print_style")+ print_style_stub.PrintStyle = lambda *args, **kwargs: types.SimpleNamespace(+ print=lambda *a, **k: None+ )+ monkeypatch.setitem(sys.modules, "helpers.print_style", print_style_stub)++ ws_stub = types.ModuleType("helpers.ws")+ ws_stub.NAMESPACE = "/test"+ monkeypatch.setitem(sys.modules, "helpers.ws", ws_stub)++ ws_manager_stub = types.ModuleType("helpers.ws_manager")+ ws_manager_stub.ConnectionNotFoundError = RuntimeError+ ws_manager_stub.get_shared_ws_manager = lambda: types.SimpleNamespace(+ emit_to=lambda *a, **k: None+ )+ monkeypatch.setitem(sys.modules, "helpers.ws_manager", ws_manager_stub)++ ws_runtime_stub = types.ModuleType("plugins._a0_connector.helpers.ws_runtime")+ ws_runtime_stub.clear_pending_computer_use_op = lambda *args, **kwargs: None+ ws_runtime_stub.computer_use_metadata_for_sid = lambda *args, **kwargs: {}+ ws_runtime_stub.select_computer_use_target_sid = lambda *args, **kwargs: "sid"+ ws_runtime_stub.store_pending_computer_use_op = lambda *args, **kwargs: None+ monkeypatch.setitem(+ sys.modules,+ "plugins._a0_connector.helpers.ws_runtime",+ ws_runtime_stub,+ )++ sys.modules.pop("plugins._a0_connector.tools.computer_use_remote", None)+ return importlib.import_module("plugins._a0_connector.tools.computer_use_remote")++def test_skills_tool_accepts_action_alias_for_search(monkeypatch, tmp_path: Path):module = _load_skills_tool(monkeypatch, tmp_path)tool = module.SkillsTool(@@ -559,6 +603,10 @@ def test_computer_use_remote_is_runtime_checked_standard_tool():project_root/ "plugins/_a0_connector/skills/host-computer-use/SKILL.md").read_text(encoding="utf-8")+ macos_skill_text = (+ project_root+ / "plugins/_a0_connector/skills/host-computer-use-macos/SKILL.md"+ ).read_text(encoding="utf-8")assert standard_prompt_path.exists()assert not (@@ -569,5 +617,36 @@ def test_computer_use_remote_is_runtime_checked_standard_tool():assert "not scoped to a single chat context" in standard_prompt_textassert "checked when the tool runs" in standard_prompt_textassert "visual verification is unavailable" in standard_prompt_text+ assert "host-computer-use-macos" in standard_prompt_textassert '"tool_name": "computer_use_remote"' in skill_text+ assert '"tool_name": "computer_use_remote"' in macos_skill_text+ assert "Backend-specific macOS guidance" in macos_skill_textassert "Beta desktop control" in skill_text+++def test_computer_use_remote_start_session_reports_backend_features_and_macos_skill(monkeypatch):+ module = _load_computer_use_remote_tool(monkeypatch)+ tool = object.__new__(module.ComputerUseRemote)++ message = tool._extract_result(+ "start_session",+ {+ "ok": True,+ "result": {+ "session_id": "s1",+ "width": 1920,+ "height": 1080,+ "backend_id": "macos",+ "backend_family": "macos",+ "features": [+ "accessibility-tree-snapshot",+ "accessibility-structural-targeting",+ ],+ },+ },+ )++ assert "session_id=s1" in message+ assert "backend=macos/macos" in message+ assert "features=accessibility-tree-snapshot, accessibility-structural-targeting" in message+ assert "host-computer-use-macos" in message