{"id":48049,"plugin_id":"plugin_asdk_app_6acd1aa56f948191ab40d2bdb1fd5050","kind":"skill","collection_source":"plugin_package","comparison_source":null,"observed_at":"2026-10-24T13:35:20.608Z","digest":"b6da35a7ad1b6c13e1a125ec506f9372124c95467ae00c722a6e73d850ebe97a","against":null,"payload":{"description":"Open Revyl Atlas maps and existing device viewers inside a compatible host, or manage an explicitly authorized device session through connected Revyl workspace tools. Use revyl-cloud-app for cloud builds and exploration, or the bundled CI-proof skill for existing artifacts.","included_files":[],"name":"revyl-workspace","skill_md_contents":"---\nname: revyl-workspace\ndescription: Open Revyl Atlas maps and existing device viewers inside a compatible host, or manage an explicitly authorized device session through connected Revyl workspace tools. Use revyl-cloud-app for cloud builds and exploration, or the bundled CI-proof skill for existing artifacts.\n---\n\n# Revyl Workspace\n\nUse the connected Revyl tools for workspace data and device sessions. This\nworkflow does not need a shell. Check the tools actually available in the host;\nnever invent a tool or silently fall back to a shell command. If the connection\nis absent or expired, ask the user to connect Revyl through the host's normal\nauthentication flow. Never request API keys, passwords, or login codes in chat.\n\n## Choose one execution path\n\n- For Atlas or a viewer, use this workflow. Opening a view is not authorization\n  to provision a device, build an app, or run a test.\n- For an existing test, use the connected test tools when available; no shell\n  is needed to run a saved test against a known build.\n- For a new app, standalone remote build, new persisted test, exploration, or other\n  CLI operation, use `revyl-cloud-app` in a shell-capable cloud workspace.\n- For a continuous development loop in an existing repository, use the bundled\n  `revyl-codex-dev-loop` skill in a shell-capable host.\n- For exact-SHA proof of an existing CI artifact, use the bundled\n  `revyl-codex-proof-ci` skill and its bounded artifact lookup.\n\nThe CLI and MCP can display the same server-issued session ID, but one workflow\nmust own its creation and cleanup. Never call `start_device_session` to recreate\na CLI-owned session just to embed its viewer. Account mismatches require the\ncorrect account connection, not a replacement session or a copied credential.\n\n## Open Atlas or an existing device\n\nCall `open_revyl_workspace` with `view=\"atlas\"` and a known authorized `app_id`.\nWhen no app is selected, use empty arguments to open the Atlas app picker. Do\nnot guess an identifier or use a URL in an identifier field.\n\nFor a device, reuse a session ID from the current workflow or resolve an exact\nsession with `list_device_sessions`. Call `open_revyl_workspace` with\n`view=\"device\"` and `session_id`. With no session selected, the device view\nopens the session list; it does not start a new device. If several sessions\nmatch, ask which one instead of choosing arbitrarily.\n\nThe tool links a client-rendered UI resource. A returned URL is not proof that\nthe iframe loaded, and the absence of an image in the text response is not\nproof that embedding failed. Report the confirmed tool outcome without\nclaiming to see the live UI. Use a normal viewer link as the fallback when the\nhost cannot render the component. Never publish or share the session merely\nto make embedding work. In ChatGPT, present the returned `chatgpt_url` as an\n**Open in ChatGPT** link. Clicking it navigates the persistent panel to the\nselected view; another tool call alone need not update an already-open panel.\nDo not invent this link when it is absent. Keep `workspace_url` as the ordinary\nbrowser fallback. Say that the view is ready to open, not that it is already\nvisible or playing, unless separate evidence confirms that outcome.\n\n## Start a device only when requested\n\nRequire an authorized app, environment, platform, and device operation. Resolve\nthe target from the user's request and available read-only tools; ask when it\nis ambiguous. Supply one artifact selector to `start_device_session`: a known\n`build_version_id` for a pinned artifact, or `app_id` for the expressly selected\napp's latest build. Use `no_open=true` so the client owns viewer navigation and\nset a bounded `idle_timeout`, normally 900 seconds.\n\nRecord the returned session ID and index. Open that same session's workspace\nwhen needed; do not invoke startup again to retry UI rendering. If startup\ntimes out, inspect `list_device_sessions` before considering another start.\nNever automatically retry an ambiguous provisioning result.\n\nWhen the request includes a later test against the same build, resolve the\nlatest compatible build once with the available build-list tool and retain its\nbuild version ID before startup. Pin both the session and the test to that ID;\ndo not resolve \"latest\" again between operations.\n\nPass the server-issued `session_id` to every session-scoped tool that declares\nit. For an older tool accepting only `session_index`, first confirm that the\nindex still maps to the intended ID. Use the tool's declared arguments for\nscreenshots, interactions, and validation. Inspect returned screenshots before\ndescribing them, and treat missing or false validation results as unverified\nor failed.\n\nFor a bounded app walkthrough, use screenshots and the declared interaction\ntools on that same session. Respect the authorized action/time budget and\navoid destructive or externally visible actions unless separately authorized.\nSeparate observed failures, usability judgments, and measured timings.\n\nTo run an explicitly requested existing test, resolve its exact name or ID.\nUse `manage_tests(action=\"run\", params={test_name, build_version_id})` in the\ncore/full profiles, or `run_test` when that tool is exposed. Inspect the returned\nstatus and poll the declared status tool within a bound when necessary; a\nqueued execution is not a passing test. A test execution owns its own device\nlifecycle and does not reuse the manual session. Do not create or change a test\nas a fallback when the named test is missing.\n\nWhen the authorized work ends, stop only the owned session with\n`stop_device_session(session_id=...)`; for an older index-only tool, recheck\nthe ID-to-index mapping first. Never set `all=true`. A successful\nrequest to stop is not the same as a released device: check the terminal result\nand the active-session list. Use bounded polling, and report pending cleanup\nwithout claiming that it completed.\n\nDo not create, modify, or publish tests, builds, workflows, shares, or reports\nunless the user requested that particular action. Keep signed media URLs,\nlaunch values, credentials, and customer content out of public artifacts.\n"},"changes":[],"summary":"First saved snapshot. No earlier version is available for comparison.","summary_kind":"deterministic","summary_metadata":{}}