菜单
Use when the user should see an agent-owned display through Intendant's shared view: demoing a finished result, letting the user watch live GUI/browser work, focusing attention on a region, capturing a display frame, or asking the user to take input authority.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | visual-collaboration |
| description | Use when the user should see an agent-owned display through Intendant's shared view: demoing a finished result, letting the user watch live GUI/browser work, focusing attention on a region, capturing a display frame, or asking the user to take input authority. |
Use this skill when work benefits from the user having live visibility into an agent-owned display — a sandbox, VM, dedicated agent machine, or virtual display: UI debugging, demos, app setup, visual inspection, or explaining what is happening on a display. The shared view streams the agent's workspace to the dashboard; it is not a mechanism for unattended control or observation of the user's personal machine.
show_shared_view: opens the shared display surface and marks the relevant display as the shared view, requesting display-stream activation so the dashboard shows it live.focus_shared_view: highlights a normalized region {x, y, width, height} on the shared display. Coordinates are fractions from 0.0 to 1.0.capture_shared_view_frame: captures the current display as an MCP image and foregrounds the same dashboard view.request_shared_view_input: asks the user to take input authority. Input authority is always granted by the user clicking the dashboard control — the tool only asks; it never grants.hide_shared_view: dismisses the banner and focus overlay when collaboration is done.Proactively open the shared view when the human should visually stay in the loop:
reason so the user can see what was done.request_shared_view_input; wait for the user to take control from the dashboard.Use focus_shared_view whenever you reference a specific UI area, keep notes short and concrete, and hide_shared_view when the shared visual moment is over.
Prefer display_id when known. Use display_target otherwise:
display_99, 99, or legacy :99 for an agent-owned virtual display — the primary case.user_session only when the user has explicitly opted into sharing their own screen (they asked you to work on or look at their desktop). Never default to it, and never treat auto-detection landing there as consent.The shared view is a dashboard coordination layer. For actual computer-use actions, continue using read_screen, take_screenshot, and execute_cu_actions; for archived stream frames, use list_frames and read_frame.
Use to operate a running Intendant daemon from the CLI — sessions, approvals, displays and screenshots, computer-use input, browser workspaces, audio, and federated peers (message or delegate to another machine's agent, or drive its screen directly with --peer). Prefer `intendant ctl` over broad MCP tools to keep model context small.
Keyless end-to-end smoke for operator goals on a NATIVE session — drives a mock-provider intendant over the control socket and proves set/get/notice-injection/clear/honest-failure
Keyless end-to-end smoke for the session-vitals git + cache segments — mock-provider daemon in a scripted git repo, asserts the merged session_vitals event on the control socket (branch/dirty/ahead/parity plus cache hit/TTL)
Use when investigating Intendant run history, session logs, dashboard replay/search behavior, external-agent wrapper logs, Codex/Claude backend logs, context rewind/fission sidecars, display frames, recordings, or task artifacts under ~/.intendant/logs. Teaches how to find sessions, search session.jsonl and sidecars, follow turn-file byte spans, map wrapper sessions to backend ids, and interpret what each log artifact provides.
E2E test the display recording and replay system. Launches intendant --web with recording enabled, triggers Xvfb display creation, verifies ffmpeg recording starts, segments are created and serveable, and the replay UI loads in the browser. Asserts via /recordings and /debug HTTP endpoints. Human monitors via VNC on port 5950.
E2E test the live audio pipeline with real audio. Uses espeak-ng TTS + PulseAudio virtual mic to feed synthesized speech to Gemini Live or OpenAI Realtime through the browser's getUserMedia on /app. Asserts via /debug endpoint JSON. Human monitors via VNC on port 5950. Tests the full audio path: TTS -> virtual mic -> Firefox -> AudioWorklet -> WASM -> live model -> tool calls -> agent.