dingtalk-aitable

Run the bundled script: python scripts/aitable_import_via_task.py <baseId> <filePath>. It requests an upload credential, PUTs the file with an empty Content-Type, and triggers the import task. For writing rows into an existing table, use scripts/import_records.py with a CSV or JSON file.

Use dws aitable +record-query with --base-id and --table-id, plus --filters as a JSON object like {"operator":"and","operands":[...]} referencing real fieldIds. Add --sort for ordering and --field-ids to return only specific columns.

Yes, but you must not paste image URLs into cells. Run scripts/upload_attachment.py <baseId> <filePath> to get a fileToken via pre-signed PUT upload, then write {"fileToken":"ft_xxx"} into the attachment field during record create or update.

AI Table is a record-based multi-dimensional database with typed fields, views, and dashboards, handled by this skill. Excel-style cell, range, and worksheet operations belong to the online spreadsheet (sheet) product, which is handled by the dingtalk-misc skill instead.

Batch shortcuts shard writes and verify each batch; partial_success means some batches failed and the checkpoint must be resumed. For unknown write results, read back by the real recordId once instead of replaying the batch, since select fields may return {id,name} objects instead of plain strings.

Reading roles works for normal members, but enabling advanced permissions or creating, updating, and deleting custom roles requires the current dws user to be a base administrator. System roles cannot be modified via CLI, and role-to-member binding must be done in the AI Table web UI.