v.26.9.26.1
This release delivers new features, new components, and enhancements across Robility Flow to improve workflow management, agent capabilities, and overall platform usability.
What’s New
New Feature
1. Skill Management: Organize and curate the reusable Skills available to Deep Agents, controlling which ones agents can draw on for discovering tools, planning multi-step tasks, and executing actions within enterprise workflows.
2. Workspace Management: Manage workspaces created by Deep Agent workflows — browse contents, edit and save files, copy paths, export content, and remove files, folders, or entire workspaces when they’re no longer needed. Every Deep Agent workflow automatically gets its own isolated workspace to store files, generated artifacts, Skills, and other resources during execution.
3. Robility Flow Dashboard: Introduced a redesigned welcome screen for Robility Flow, giving users a quick overview of their workspace at login, including active/inactive workflow counts, token usage, execution success and failure rates, and a list of recently run workflows with their status and duration.
New Components
Flow introduces new components to improve loop control, file handling, and data extraction across automation flows.
1. Do-While Loop (Beta): Establish precise control over repeated execution with condition-driven looping. This component runs a connected workflow at least once, then repeats it while the condition is True and stops when it is False, ensuring only the result from the final iteration is retained for clean and predictable outputs.
2. Merge Files (Beta): Simplify document consolidation by combining multiple files into a single PDF. This component automatically uploads the merged file to the Robility Flow File Manager, ensuring organized, accessible, and consistent file handling across your automation flows.
3. Get Transaction Count: Monitor workload status with real-time transaction visibility. This component retrieves the count of transactions with New and Locked statuses from the connected project in Interact, enabling workflows to use current counts for further processing and decision-making.
4. Multi-Field Nested Parser: Extract exactly the data you need from structured sources such as JSON, Tables, and DataFrames. By selecting multiple fields or nested values in one step, this component returns each field as an individual output that connects directly to downstream components.
Enhancements
1. Preloaded MCPs: The following MCPs are now preloaded and available for use out of the box: Gmail, Google Calendar, Google Cloud BigQuery, Databricks, Apache Spark & Slack
2. Invoke Flow: Enhanced the Invoke Workflow functionality to support invoking and running a published workflow created by another user in the same project, while maintaining backward compatibility with existing user-level workflow invocation.
3. Save File: Introduced a new Raw Content input to save user-provided content directly as a file in the configured Robility Flow default or KB file directory. The generated file path is returned, enabling downstream flows and tools to consume it seamlessly.
4. Deep Agent: Enhanced the Deep Agent component to integrate with the new centralized Skill Management and Workspace Management capabilities, exclusive to Deep Agent workflows.
a. Operates securely within its own isolated workspace, using only user-selected Skills.
b. Improved sub-task execution for deeper task breakdown and result retrieval.
c. Browse and manage all Deep Agent workspaces under a project, including removing unused ones, without storing workspace data in the database.
d. Optimized Skill and context handling to minimize unnecessary token consumption.
5. Publish Flows as Reusable Templates
Flow publishing now includes a Use as a Template option, allowing published flows to be reused as templates for creating similar workflows. This helps accelerate flow creation and maintain consistency across projects.
Bug Fix
1. Scheduler
Fixed an issue where scheduled executions could get stuck indefinitely in the Run Initiated state. Scheduling logic has been improved to process execution records correctly and update execution status as expected.
2. Robility KB Ingestion Component
Fixed an issue where invalid LLM configurations could be saved when creating or updating a Knowledge Base. LLM credentials are now validated upfront in Robility Flow, ensuring only valid configurations are used.
Known Issues
1. Scheduled Flow Execution: When a flow is scheduled with an interval of more than one hour, it may execute one additional time. (Tentative fix: Oct 2026)
2. The embedding functionality in the Robility LLM component is currently not working.( Tentative fix: Oct 2026)
3. Deleting a project also removes all associated flows, and they cannot be recovered.(Tentative fix: Oct 2026)
4. If the description field is removed from the flow JSON before import, the workflow is still uploaded with an empty description. (Tentative fix: Oct2026)
5. Flow names are user-specific, allowing multiple users to create flows with the same name, which may result in duplicate solution names and versions. (Tentative fix: Oct 2026)
6. Newly generated API keys currently apply to all workflow versions; this will be updated to restrict keys to specific versions in a future release.(Tentative fix: Oct 2026)
7. The Web Scraper component does not support extracting data from JavaScript-based web pages. (Tentative fix: Oct 2026)
Existing Limitations
1. A user cannot edit or schedule a flow that has already been scheduled by another user within the same project.
2. The Model Provider integration is currently accessible only to the user who configured it. Other users within the same project or agent cannot access or use the configured integration.
Workaround: Use the LLM component by passing the required LLM connection as an input. This enables other users to utilize the LLM functionality without requiring direct access to the configured Model Provider integration.
3. The Pulse Component does not support variable mapping for the Security Key and Public Key input fields when dynamically loading prompts.
Workaround: The component works as expected when the values are provided directly instead of using variable mapping.
4. Version History will track only workflows published after the upgrade. Previously published versions remained available in Robility Manager but will not appear in Version History or be eligible for scheduling. To enable these features, export and rename the workflow (for example, FlowName.1.0.0 → FlowName), then republish it.
5. The Loop component has been enhanced to resolve stale and duplicate Datatable updates. Updating the component will remove all nodes within the Loop component and associated Datatable values. Back up the workflow before upgrading.
6. When processing large files with the Read File component, increase the Timeout and Retry Count settings to ensure successful OCR extraction.
7. The invoke flow in tool mode does not work with the configured slug name.
Workaround: After connecting the tool node, manually update the slug name before using it.
Release Date: 26.09.2026
