create_custom_object
ChatGPTCreate a new custom object record of a given leaf_type. Call discover_schema(leaf_type='<TYPE>') first to learn the available custom_fields for this type.
create_object
ChatGPTCreate a new DevRev object using the specified create action (e.g., create_issue, create_ticket). IMPORTANT: Always call discover_schema(action_name='create_<object_type>') first to get the required fields, optional fields, and valid enum values before calling this tool. Pass the action_name and a values object containing the field key-value pairs. Example: create_object(action_name='create_issue', values={'title': 'Fix login bug', 'applies_to_part': 'PROD-123', 'priority': 'p1'}).
discover_schema
ChatGPTRetrieve the JSON input schema for a specific DevRev action, or list all available actions when action_name is omitted. Call this tool BEFORE calling any other action tool to learn the required and optional parameters, supported object types, and valid enum values. Examples: discover_schema() → lists all actions; discover_schema(action_name='create_issue') → returns the full input schema for creating an issue. discover_schema(action_name='create_issue', subtype='bug_report') → returns the full input schema for creating an issue with the subtype 'bug_report'. discover_schema(leaf_type='BOOK') → returns the custom_fields schema for create_custom_object / update_custom_object calls on the 'BOOK' custom type. discover_schema(leaf_type='BOOK', subtype='form_two') → also includes subtype-specific (ctype__) fields. IMPORTANT: Always call this tool first with no arguments to list all available actions before calling any other action tool ie. discover_schema() -> returns all available actions with descriptions (eg. create_issue, update_issue, list_issues, etc.).
download_artifact
ChatGPTGet a presigned download url for the provided Artifact ID. This does not return the bytes — issue a GET to the returned url to download the artifact.
execute_dql
ChatGPTExecute a DQL (DevRev Query Language) query to search and filter DevRev objects. Returns matching objects with pagination cursors (next_cursor / prev_cursor). Supported object types: issues, tickets, enhancements, articles, accounts, rev_users, custom_objects, conversations, incidents, meetings, tags. Resolving a DevRev app URL into a DQL query When the user shares a URL like https://app.devrev.ai/.../views/..., follow these steps: 1. DECODE & PARSE THE URL - URL-decode the full string. - Extract the widget DON from the path segment after "views/" (e.g. don:.../abc). - Record any URL query params that represent active filter overrides: object_type, referenceId, custom_stage, space_id, and any *_op suffix params. 2. FETCH THE CONTAINER'S CONTEXT - Call fetch_object_context(<container DON>). - Widget → inspect sub_widgets[]; each entry has a title, reference_id, and dql_query.filter. Match referenceId from the URL to the correct sub_widget to get the active tab's filter. - The fetch_object_context result contains the dql_query to use directly. 3. EXECUTE THE QUERY - Use the dql_query.filter from fetch_object_context as the filter input to this tool. - Apply any active filter overrides from the URL params on top of that base filter. Example { "dql_query": { "filter": { "op": "and", "operands": [ {"field":"enhancement.stage_v2.stage","op":"in","value":["don:.../73","don:.../5","don:.../20"],"value_type":"json"}, {"field":"enhancement.object_type","op":"in","value":["enhancement"],"value_type":"json"}, {"field":"enhancement.space","op":"in","value":["don:.../EUsdg1Sw"],"value_type":"space","include_child_spaces":true,"include_parent_spaces":false} ] }, "limit": 50 }, "fields": ["id", "title"] }
fetch_object_context
ChatGPTRetrieves rich contextual information about any DevRev object. Accepts display IDs (ISS-123, TKT-456, INC-789, ENH-101, etc.) and full DON IDs (don:core:dvrv-us-1:devo/0:issue/123) equally — no need to resolve one to the other first. Returns a detailed summary including current state, linked objects, timeline history, and relevant metadata. Use this before updating, linking, or triaging an object to understand its current context.
get_project
ChatGPTGets the project details for the given project ID
get_self
ChatGPTRetrieves details of the current user from the context
get_timeline_entries
ChatGPTLists timeline entries for a DevRev object, newest first by default. Timeline entries include comments, system events, and activity history. Required: object ID (ticket, issue, conversation, etc.). Optional: created_by (author DON IDs — matches any provided ID; pass your user or service account ID to find your own comments), entry_type (comments or events), sort_order (newest_first or oldest_first), limit, and date filters. System user entries require include_bot_messages=true. For email comments, body can be a truncated preview (often cut at 1000 characters); set include_html_body=true to also get the full email as HTML in html_body; html_body can be large, so pair it with a small limit (for example 10). Use this to review an object's interaction history or find entries by a specific author.
get_valid_stage_transitions
ChatGPTUse this tool when transitioning the stage of an object. Returns the valid stage transitions for a given leaf/object type and its current stage.
hybrid_search
ChatGPTPerforms full-text and semantic search across DevRev's knowledge graph. Required: query (1–400 chars), namespace (object type to search). Optional: limit (1–50, defaults to 10). Use this to discover object IDs by name, find related items, or explore what exists before creating or linking. Common namespaces: issue, ticket, enhancement, article, account, conversation, incident, opportunity, vista (sprint boards), dev_user (internal team members), rev_user (external customer contacts), product, capability, feature, tag, group, dashboard, meeting, widget, question_answer (search FAQ/Q&A content), directory, dm, custom_object.*.
link_objects
ChatGPTCreate a link between two DevRev objects, read as 'source <link_type> target'. Pass the source and target display IDs (e.g. 'ISS-123', 'ENH-456') and the link_type describing the relationship; direction matters, so order source and target to match. The link is validated against the object types involved, and an error is returned if the link_type is not allowed between them.
list_artifacts
ChatGPTList DevRev-hosted artifacts attached to an object and its comments. Returns artifact IDs, file names, MIME types, sizes, and the source object or comment. Use download_artifact with a returned ID to get a signed download URL.
list_objects
ChatGPTList DevRev objects using the specified list action (e.g., list_issues, list_tickets) with optional filters. IMPORTANT: Always call discover_schema(action_name='list_<object_type>') first to get the available filter parameters, sortable fields, and valid enum values. Pass filters in the values object and optionally restrict returned fields with the fields array. Example: list_objects(action_name='list_issues', values={'applies_to_part': ['PROD-123'], 'priority': ['p0','p1']}, fields=['display_id','title','priority','stage']).
mark_notifications
ChatGPTMark inbox notifications with a specific state (e.g., read, unread). Provide notification IDs and the desired state to update.
show_notifications
ChatGPTShows the user's inbox notifications grouped by parent object (e.g., all notifications about the same issue together). Each group shows the parent object name when available, plus a list of notifications with title, body, type, and ID. Supports filtering by state (read/unread), priority (high/medium/low), and date range (since/until as RFC3339, or last as a relative duration like 24h, 7d). Use pagination (page/per_page) to navigate groups and per_group to limit notifications within each group.
update_custom_object
ChatGPTUpdate an existing custom object record. The 'id' (DON) of the object is required. Call discover_schema(leaf_type='<TYPE>') first to learn which custom_fields are available.
update_object
ChatGPTUpdate an existing DevRev object using the specified update action (e.g., update_issue, update_ticket) by providing its ID and the fields to modify. IMPORTANT: Always call discover_schema(action_name='update_<object_type>') first to get the updatable fields, their types, and valid enum values. The 'id' field is always required inside values. Only include fields you want to change — omitted fields remain unchanged. NOTE: When updating any custom type field (prefixed with 'ctype__'), you must also pass the 'subtype' parameter in the request — omitting it will cause the update to fail. Example: update_object(action_name='update_issue', subtype='my_custom_subtype', values={'id': 'ISS-123', 'ctype__custom_field': 'value'}).