{"id":95920,"plugin_id":"plugin_asdk_app_69fb51f9519081919c1f3e44ea9a5a05","kind":"skill","collection_source":"plugin_package","comparison_source":null,"observed_at":"2026-11-14T00:15:18.932Z","digest":"fdade6740622fd6e111596c241b383c66c714812bfa427afcdc572971ec4a41e","against":null,"payload":{"description":"Use for Deepnote workspace and project operations: summaries, project and folder inventories, project inspection, integrations and their usage, cached table structure, and reading, writing, copying, or deleting project files.","included_files":[],"name":"deepnote-workspace","skill_md_contents":"---\nname: deepnote-workspace\ndescription: \"Use for Deepnote workspace and project operations: summaries, project and folder inventories, project inspection, integrations and their usage, cached table structure, and reading, writing, copying, or deleting project files.\"\n---\n\n# Deepnote Workspace\n\nTool arguments, pagination, and the capability boundary are defined in `deepnote-mcp`. Build links with `deepnote-links`.\n\n## Workspace Summary Workflow\n\nWhen the user asks for a workspace summary, heartbeat, overview, or which notebooks are active or scheduled:\n\n1. Use `get_me` for workspace name, workspace ID, API key type, and caller access level when useful.\n2. Use `list_projects` to collect projects and notebooks. For complete inventories, page through all results.\n3. Use `list_integrations` to collect integration names, types, and IDs.\n4. Use `get_notebook` for notebooks that need block, input, or recent-run detail.\n5. Identify scheduled notebooks from `isScheduled`.\n6. Identify active notebooks from `lastRunAt` or an explicitly requested run status from `get_run`. `get_machine` reports the project machine's state, not open editor sessions. Say that active means recent run activity rather than an open editor session. Prefer `recently run`, `scheduled`, `pending run`, or `last run` over \"currently open\" or \"currently running\".\n7. Map integration usage with the usage tools when direct usage matters. If usage was not checked, write `Usage not checked`; if a checked usage tool returns nothing, write `None found`.\n8. Build project and notebook links with `deepnote-links`, using `utm_term=workspace_summary` for links created by this workflow rather than by a single tool result.\n\n## Workspace Summary Output\n\nWorkspace-status output should read like a small operations dashboard:\n\n1. Start with a one-sentence health line, for example: `Deepnote workspace is reachable; the current MCP response includes 6 projects, 15 notebooks, 1 scheduled notebook, and 4 integrations.`\n2. Add a compact `Key Signals` list with counts visible in the current MCP response: projects, notebooks, scheduled notebooks, recently run notebooks, failed or pending runs when checked, and integrations.\n3. Use the summary table below as the main artifact: one row per notebook, grouped by project.\n4. Keep integrations inside the table as an `Integrations` column. Use a standalone integration table only when the user explicitly asks for an integration inventory or usage report.\n5. Hyperlink project and notebook names. The notebook name is the link label.\n6. Finish with `Notable Findings` only when something is actionable: a scheduled notebook with no last run, a pending or failed run, a notebook that prints environment variables, or an integration with no checked usage.\n\nSummary table, for workspace summaries, notebook inventories, project breakdowns, and \"which notebooks do I have?\" requests:\n\n| Project | Notebook | Scheduled | Last Run Seen | Integrations |\n| --- | --- | --- | --- | --- |\n| [Project name](project URL with UTM parameters) | [Notebook name](notebook URL with UTM parameters) or `N more notebooks` | `Yes` or `No` | `YYYY-MM-DD HH:MM UTC`, `None seen`, or `Not visible via MCP` | `Integration name/Type` or `None found` |\n\nRow rules:\n\n- For a specific project or notebook, filter the table to it.\n- For large workspaces, keep individual rows for scheduled notebooks, recently run notebooks, and notebooks with visible integrations, then collapse each project's remaining notebooks into one `N more notebooks` row. A collapsed row is `No` for Scheduled, shows the most recent visible `lastRunAt` among its notebooks or `None seen`, and lists their integrations, `None found` when usage was checked, or `Usage not checked`.\n- `Last Run Seen` is the notebook's visible `lastRunAt`, or a checked `get_run` completion time when more current. Format dates in UTC as `YYYY-MM-DD HH:MM UTC`. Never write `None seen` when a run ID or run timestamp is visible.\n- `Integrations` uses names and IDs from `list_integrations`, mapped with the usage tools when direct usage matters. Visible integration references from inline run snapshots may also be mentioned. Do not infer usage from integration names alone; write `None found` only when checked usage or visible references return no connection.\n\n## Project And Folder Inspection\n\nUse `get_project` for a specific project's type, containing folder, notebooks, attached integration summaries, recursive file inventory, data app settings, and environment. Use `get_machine` for whether its machine is running and how big it is. Use `get_notebook` only when block or input detail is needed. Report the relevant notebooks and integrations, a file count unless individual paths were requested, and whether a data app is published.\n\nTo put a project on another image, resolve the environment with `list_environments` and pass its `id` as `environmentId` to `create_project` or `update_project`. Switching the environment of an existing project restarts its running machine, which clears kernel state.\n\nUse `list_folders` to resolve the `folderId` for creating or moving a project, and create a missing folder with `create_folder`. Folder names are not unique, so resolve nested paths by walking `parentFolderId`; omit `nameContains` when the complete hierarchy is needed.\n\n## Integration Mapping Workflow\n\n1. Use `list_integrations` to resolve an integration name or type to an ID.\n2. Use `get_integration` for cached structure. The response includes integration details plus `tablesJsonPreview`, the serialized matching cached table structure, and `tablesTruncated`. Parse `tablesJsonPreview` as JSON only when `tablesTruncated` is false. When it is true, use `tableNames` to choose narrower `databaseName`, `schemaName`, or `tableName` filters, or describe the tables needed in `search`, and call `get_integration` again. A preview of `[]` means no matching cached structure is visible through MCP, not that the live database lacks the table.\n3. Use `list_integration_project_usages` for connected projects, `list_integration_notebook_usages` for notebooks with SQL blocks that use the integration, and `list_integration_block_usages` for the exact SQL blocks and their content. Narrow any of them with `projectId`.\n4. Say that structure is cached. Do not present it as a live database scan, and do not claim access to row previews, query results, file metadata, or environment configuration unless an exposed MCP tool or a run snapshot provided them.\n\n## Integration Management\n\n1. Resolve the integration with `list_integrations` and inspect the project's current attachments with `get_project`.\n2. Use `attach_integration` or `detach_integration` only when the user asked to change that project. These operations return a conflict when the integration is already in the requested state; report the existing state instead of retrying.\n3. Use `create_integration` only when the user wants a new workspace integration and its type is accepted by the current tool schema. It requires integration-management permission and creates the integration without attaching it to a project.\n4. The creation response repeats connection `metadata`, which may include credentials. Report only the new integration's name, type, and ID.\n\n## Cross-Project File Copy\n\nUse `get_project` on both projects to confirm the source file exists and the target path is free. `copy_file` preserves the normalized source path, accepts no destination path, requires different source and target projects, copies files rather than directories, and never overwrites. It requires read access to the source and edit access to the target.\n\nStandalone integration table, only when explicitly requested:\n\n| Integration | Type | Visible Notebook Usage |\n| --- | --- | --- |\n| `Integration name` | `type` | `Project / Notebook` from usage tools, `None found`, or `Usage not checked` |\n\n## Project File Reads, Writes, And Deletes\n\n1. Use `read_file` only for small text-based files, never for large or binary files. Pass `startLine` and `endLine` to read a long file in parts while `hasMore` is true.\n2. Use `write_file` only for small text-based files, such as a script, config, or CSV, never to upload large files or big binary blobs. Check the project's file inventory with `get_project` first.\n3. Pass `overwrite: true` only when the user asked to replace an existing file. Without it, a taken path gets a generated unique path, so report the returned `file.path`.\n4. Use `delete_file` only for a file the user named, after confirming its path with `get_project`. It cannot be undone, and it deletes a Streamlit app that serves the file, so check `list_streamlit_apps` first.\n5. Writing and deleting require edit access to the project; reading needs View access.\n"},"changes":[],"summary":"First saved snapshot. No earlier version is available for comparison.","summary_kind":"deterministic","summary_metadata":{}}