1.2
1.2.0-funcrel
Note
This release includes a structural change to client connectivity. MCP-aware clients no longer connect directly to the MCP Server - instead, all clients must now connect directly to CAST Imaging via its public gateway. As a result, CAST Imaging 3.5.3-funcrel or later is required as the target installation.
New Features
| Summary |
Details |
| Introduced a new "Profiles" tool to retrieve high-level profiles for objects, transactions and data graphs |
The MCP Server now includes a set of Profiles tools that allow users to retrieve summarized, profile-level information across key application dimensions: 1. Object Profiles: High-level distribution and characteristics of application objects. 2. Transaction Profiles: Profiles of API and UI endpoints (transactions) to understand interaction patterns. 3. Data Graph Profiles: Profiles of data entity interaction networks (data graphs) to analyze data flow and relationships. Each tool calls the corresponding backend API to fetch profile data and returns the results in a paginated, structured format for easy consumption. These profiles enable users to quickly assess application composition, identify dominant patterns, and support architectural analysis without needing to drill into individual objects or interactions. |
| Introduced a unified "Pathfinder & Call Hierarchy" tool to analyze execution paths and object relationships |
The MCP Server now provides a powerful Pathfinder & Call Hierarchy analysis tool that consolidates two key capabilities into a single, flexible interface: 1. "Path Finder Analysis": Allows users to identify the shortest or all possible execution paths between a source object and a target object. This helps in understanding how components interact, tracing dependencies, and analyzing execution flows across the application. 2. "Call Hierarchy Analysis": Enables users to explore the caller or callee hierarchy of a specific object, revealing how it interacts with surrounding components and how responsibilities are distributed within the system. Users can control the depth and direction of analysis by adjusting parameters such as call direction (caller, callee, or both), traversal depth (hops), inclusion of external objects, and level of object granularity (main objects only or including sub-objects). The MCP Server invokes the underlying graph APIs to retrieve both hierarchy and pathfinding data and returns the results This allows users to clearly visualize dependencies, understand object interactions, and perform impact or architectural analysis efficiently. |
| Introduced a new "API Inventory" tool to fetch and explore API details |
The MCP Server now provides a new API Inventory tool that allows users to retrieve detailed information about APIs within an application. The tool supports two modes of operation: 1. "List Mode": Fetches a paginated list of APIs with optional filtering by API name, type (e.g., Public/Private), and method. 2. "Graph Mode": Retrieves the dependency graph for a specific API using its unique ID, enabling deeper analysis of call hierarchies and relationships. Results are presented in a clear, structured format, making it easier for users to analyze API usage, understand dependencies, and make informed architectural and operational decisions. |
Feature Improvements
| Summary |
Details |
| Streamlined and improved installer |
The MCP Server installer has been streamlined into a single "run.sh" script that handles both fresh installs (via "–install" and updates (via "–update"). This release also removes all legacy HTTPS/Nginx components to simplify the architecture. See the installation and update guide. |
| Enhanced the "quality_insight_violations" tool |
The MCP Server now extends the existing "quality_insight_violations" tool to better support bookmarked insights, particularly for structural flaws and related quality rules. When the "IMAGING_CODE" parameter is set to true, users can retrieve detailed occurrence information for bookmarked insights, including file names, line numbers, and relevant source code snippets. |
| Enhanced the "object_details" tool with new endpoints |
The MCP Server now supports a new endpoints focus within the existing "object_details" tool, allowing users to query and analyze the start points and end points associated with an object. This enhancement helps users quickly identify how an object is invoked (entry points) and where it leads (exit points), supporting more accurate change-impact and dependency analysis. This enhancement improves visibility into object exposure and interaction boundaries without introducing a new tool, keeping the workflow consistent while extending analytical capabilities. |
| Update to MCP architecture |
The process of connecting MCP aware clients with MCP Server has been changed in this release: all connections to the MCP Server must now be routed through CAST Imaging gateway. This change simplifies the HTTPS setup for the MCP Server, as no separate reverse proxy is required when accessing MCP over HTTPS. To enable HTTPS connectivity, CAST Imaging 3.5.3-funcrel is required and must be configured to run over HTTPS. See the installation guide. |