eosphoros-ai/DB-GPT
40.5
Weak · 5 November 2026
329.3k
lines of production code
Python
with TypeScript
4
measurements over time
What this system is
Features
Add chat mode toggle and prompt configuration form
Users can now switch between 'Preview' and 'Editor' modes when interacting with database execution or dashboard scenes via a new mode tab component. Additionally, a new prompt form allows users to configure prompts by selecting a chat scene, entering a sub-scene, defining a prompt name, and providing content, with validation for all required fields.
web/components/chat/mode-tab, web/components/prompt · high confidence
Add new colorful and status-specific icon components
The \web/components/icons\ directory now includes a set of new SVG icon components to support the unified color theme and dashboard UI updates. This adds colorful variants for navigation items (\ColorfulChat\, \ColorfulDashboard\, \ColorfulData\, \ColorfulDB\, \ColorfulDoc\, \ColorfulExcel\, \ColorfulPlugin\) and specific status indicators for file operations (\FileDone\, \FileError\, \FileSync\), as well as utility icons like \DarkSvg\ and \ModelSvg\. These are exported via the main \index.tsx\ barrel file for use throughout the frontend.
web/components/icons · high confidence
Added Alembic database migration configuration
The pilot/meta\_data area now includes the standard Alembic configuration files (alembic.ini, env.py, script.py.mako, and README) to support database schema versioning and migrations. This setup initializes the migration environment using the existing database manager, enabling automated tracking of schema changes.
_pilot/meta\data · high confidence
Added Chinese (Simplified) localization for chat, knowledge, and observability features
Users can now see the interface in Simplified Chinese. This change introduces new translation files for the chat interface (including multi-file upload, code preview, and replay features), knowledge base management (including external data source bindings for Feishu and Yuque, and LLM-Wiki generation), workflow editor, and the observability dashboard (agent traces and metrics).
web/locales/zh · high confidence
Added Docker build automation and MySQL configuration example
The Docker directory now includes a \build\_all\_images.sh\ script that automates the build process by first invoking the base image builder and then the all-in-one image builder, ensuring the base build succeeds before proceeding. Additionally, an example MySQL configuration file (\my.cnf\) has been added to the examples directory, providing default settings for character sets (utf8mb4) and authentication plugins to support database deployments.
docker · high confidence
Added DuckDB-to-MySQL and DuckDB-to-SQLite migration scripts
New example scripts have been added to the metadata directory to facilitate migrating chat history and connection configuration data from DuckDB databases to MySQL and SQLite. Users can now easily transfer their existing metadata to these alternative storage backends using the provided \duckdb2mysql.py\ and \duckdb2sqlite.py\ utilities.
docker/examples/metadata · high confidence
Added GraphRAG and Transformers lore test data
Added new test files in the examples/test\_files directory to support GraphRAG and data application testing. The graphrag-mini.md file provides structured entity and relationship data for the TuGraph and DB-GPT project ecosystems, including developer connections and organizational affiliations. The graphrag-test.md file contains comprehensive documentation for the DB-GPT framework, detailing its architecture (RAG, GBI, Multi-Agents, AWEL), supported models, and usage guides. Additionally, tranformers\_story.md was added, containing a detailed chronological narrative of the Transformers universe lore, likely serving as a test case for long-context processing or narrative generation capabilities.
_examples/test\files · high confidence
Added Windows and Linux scripts to load example SQL data
Users can now load example SQL data into SQLite databases on both Windows and Linux/macOS systems using the new \load\_examples.bat\ and \load\examples.sh\ scripts in the \scripts/examples\ directory. These scripts automate the execution of SQL files (defaulting to \\\sqlite.sql\ and \\\_sqlite\_wide.sql\) into specified database files, supporting custom paths via command-line arguments and requiring the \sqlite3\ command-line tool to be installed.
scripts/examples · high confidence
Added model icons for Bailing, Hugging Face, and Yi
The web interface now displays distinct visual icons for the Bailing, Hugging Face, and Yi models. New SVG assets have been added to the public models directory to provide better visual identification for these specific model providers within the application.
web/public/models · high confidence
Added sample data scripts for MySQL and SQLite dashboards
New Python scripts (\test\_case\_mysql\_data.py\ and \test\_case\_sqlite\_data.py\) have been added to the dashboard example directory. These scripts generate and insert sample data into \user\ and \transaction\_order\ tables for both MySQL and SQLite databases, enabling users to quickly populate their local environments with test data for dashboard visualization.
docker/examples/dashboard · high confidence
Added sample datasets for Excel examples
The \docker/examples/excel\ directory now includes two CSV files (\Walmart\_Sales.csv\ and \zx.csv\) to serve as sample data for the Excel example. \Walmart\_Sales.csv\ provides weekly sales data with associated economic indicators, while \zx.csv\ contains a renovation budget breakdown.
docker/examples/excel · high confidence
Added wiki revision diff utilities
New utility functions have been added to the wiki module to support line-level and field-level diffing of wiki content. The \line-diff.ts\ module provides a dynamic-programming based LCS algorithm to compare text line-by-line, handling additions, deletions, and unchanged lines, with a safeguard that degrades to block-level diffs for large inputs. The \revision-diff.ts\ module uses this to compare two wiki revision snapshots, returning a structured list of changed fields (title, summary, content) with their respective line diffs, omitting any fields that remain unchanged. Unit tests in \wiki-diff.test.ts\ verify the correctness of these diffing behaviors.
web/utils/wiki · high confidence
French localization for CLI, core, and app components
Added automatically generated French translation files (\.po\) for the CLI, core, app, and extension modules. This enables French language support for user-facing strings in the command-line interface, agent operators, configuration settings, knowledge graph features, and data source connectors.
i18n/locales · high confidence
Initial database schema definition for Dbgpt
The \assets/schema/dbgpt.sql\ file has been added, establishing the foundational metadata database structure for the application. This schema defines the core tables required for operation, including \knowledge\_space\ and \knowledge\_document\ for managing RAG knowledge bases, \chat\_history\ and \chat\_history\_message\ for storing conversation logs, \connect\_config\ for database connection settings, and tables for plugin management (\my\_plugin\, \plugin\_hub\) and prompt handling (\prompt\_manage\). It also includes the \alembic\_version\ table to support database migration tracking.
assets/schema · high confidence
Initial release of DB-GPT-Web frontend
This change introduces the initial codebase for the DB-GPT-Web frontend, a Next.js-based chat UI. It includes configuration for ESLint, Prettier, Tailwind CSS, and TypeScript, along with a Next.js setup that proxies API requests to the backend. The UI features a model configuration page for managing LLM providers, an agent marketplace for installing and managing plugins, and a Monaco-based SQL editor enhanced with OceanBase syntax support.
web · high confidence
Introduce SKILL mechanism and new analysis/report skills
Adds a new SKILL mechanism to the DB-GPT agent framework, allowing agents to load and manage predefined skill packs for modular and reusable capabilities. This includes an integration guide for implementing the SKILL system, along with new skills for structured business report generation (using template-driven, traceable data analysis), CSV/Excel deep data analysis (with automated statistical summaries and ECharts visualizations), headless browser automation, and specific visualization styles (G2 charts, brutalist design).
skills · high confidence
Introduce agent-based benchmark evaluation support
The models evaluation interface now supports creating and managing benchmarks for both LLM and Agent evaluation types. Users can select the evaluation type (LLM or Agent) when creating a new task; for Agent evaluations, the new modal allows configuring API endpoints, HTTP methods, headers, and response parsing strategies (JSON Path or Direct). The evaluation list displays these tasks with their specific environment (DEV/TEST) and status, enabling users to monitor and view details of agent-based benchmark runs alongside standard model evaluations.
_web/components/models\evaluation · high confidence
Introduce all-in-one Docker image with bundled MySQL
Users can now deploy a self-contained application image that includes a pre-configured MySQL database, eliminating the need for external database setup. The new Dockerfile installs MySQL, configures it for remote access with UTF-8 support, and initializes the schema via entrypoint scripts, while helper scripts provide convenient commands to build and run the container with default environment variables.
docker/allinone · high confidence
Introduce dbgpt-app package with CLI and build infrastructure
This change introduces the \dbgpt-app\ package (version 0.8.2), providing a dedicated CLI entry point (\dbgpt webserver\) for starting the DB-GPT web server. The package includes a custom Hatch build hook (\hatch\_build.py\) to dynamically resolve and bundle builtin skills, example files, and pilot templates into the wheel for both editable and standard installations. It also establishes the core application initialization logic, including database storage setup, component registration (embedding, rerank, scheduler, AWEL), and static file mounting for the web UI.
packages/dbgpt-app · high confidence
Introduce dbgpt-client Python SDK with CLI and API bindings
The dbgpt-client package is introduced as a new Python SDK for interacting with the DB-GPT API (version 0.8.2). It provides a \Client\ class for programmatic access to chat completions, AWEL flows, knowledge spaces, datasources, and evaluations, along with a new command-line interface (CLI) for running AWEL flows and managing apps. The SDK defaults to the v2 API endpoint and includes specific fixes for HTTP method usage, ensuring that \create\_datasource\ and \create\_flow\ correctly use POST requests instead of GET, and that \update\_flow\ includes the flow UID in the PUT path.
packages/dbgpt-client · high confidence
Introduce dbgpt-core package with v0.8.2 version and multi-layer agent context management
The \packages/dbgpt-core\ location now contains the foundational \dbgpt-core\ package (version 0.8.2), which provides the core \SystemApp\ and \BaseComponent\ infrastructure, lazy-loading for sub-libraries (core, rag, model, agent, etc.), and a centralized configuration system. A key behavioral addition is the multi-layer context management system for agents, which automatically compacts conversation history to prevent token overflow through four progressive layers: truncating old tool outputs, dropping old conversation rounds, generating LLM-based summaries, and emergency trimming. This system is integrated into the \ConversableAgent\ base class, allowing long-running ReAct sessions to maintain context within budget limits.
packages/dbgpt-core · high confidence
Introduce dbgpt-serve package with agent orchestration and analytics capabilities
The new dbgpt-serve package (v0.8.2) provides the core serving layer for multi-agent applications, featuring an App Manager that orchestrates conversational agents and persists their memory to a database. It includes a dedicated API router for multi-agent chat interactions and a suite of expandable analytics agents—such as Anomaly Detection, Volatility Analysis, and Report Generation—along with their corresponding execution actions and visualization protocols.
packages/dbgpt-serve · high confidence
Introduces TypeScript type definitions for the web frontend
Adds a comprehensive set of TypeScript type definitions across the \web/types\ directory to provide strict typing for the frontend application. These new files cover core domain models and API contracts for agents, applications, chat interactions, database connections, flow editors, knowledge management, model providers, prompt templates, scheduled tasks, sub-agents, and user information, ensuring type safety and better developer experience for the web UI.
web/types · high confidence
Introduces auto-chart component with custom multi-measure charts and image download
The new \AutoChart\ component in \web/components/chart/autoChart\ enables automatic visualization recommendations by integrating the \@antv/ava\ advisor with custom chart definitions. It introduces three new chart types—\multi\_line\_chart\, \multi\_measure\_column\_chart\, and \multi\_measure\_line\_chart\—which support multiple measures on a single axis and handle data sorting and encoding. The component also merges model-suggested chart types with AVA recommendations, provides a UI to switch between recommended views, and adds a button to download the current chart as a PNG image.
web/components/chart/autoChart · high confidence
Introduces dbgpt-ext package with connector catalog and database dialects
The new dbgpt-ext package (v0.8.2) provides a centralized registry for external integrations and database connectors. It includes a connector catalog (catalog.json) defining MCP server integrations for platforms like Feishu, DingTalk, GitHub, Notion, and Tavily, alongside SQLAlchemy dialects for OceanBase, StarRocks, and Vertica to enable direct database connectivity.
packages/dbgpt-ext · high confidence
Introduces new web application components and backend infrastructure
This update adds several new frontend components, including a reusable MenuModal, a default chat interface with app recommendations, and a modal for creating and editing applications. It also introduces a new evaluation page for managing datasets and evaluation metrics. On the backend, the web application now includes a database layer using Sequelize with MySQL for storing user and feedback data, along with utilities for session management via iron-session, standardized API responses, and Google One Tap authentication.
(repo-wide) · high confidence
Introduction of chat context management and internationalization infrastructure
This change introduces the foundational infrastructure for the web application's chat interface. It adds a new \ChatContextProvider\ that centralizes state management for chat sessions, including theme mode (light/dark), model selection, dialogue history, and admin lists, while automatically refreshing the available model list when changes occur elsewhere in the app. Additionally, it establishes the internationalization (i18n) setup using \react-i18next\ with support for English and Chinese locales, providing the necessary hooks and types for localized UI text.
web/app · high confidence
Native AI application framework with agent and flow support
The web interface now supports creating and editing native AI applications. Users can configure agents with LLM strategies and manage dynamic resources via the new AgentPanel and ResourceCard components. The AppModal provides a unified interface for app creation, supporting both single-agent and multi-agent (auto\_plan) modes, as well as DAG-based flow applications via the DagLayout component. Existing apps can be published, unpublished, collected, or deleted through the updated AppCard.
web/components · high confidence
New API client module for application management and configuration
A new API client module has been introduced at \web/client/api/app/index.ts\ to handle frontend interactions with the backend application service. This module exposes functions for managing application lifecycles (creating, updating, and listing apps), retrieving configuration data such as team modes, available agents, and LLM strategies, and managing application-level permissions (admins). It also includes endpoints for fetching native app scenes and resource parameters, providing a centralized interface for the frontend to interact with the \/api/v1/app\ and related endpoints.
web/client/api · high confidence
New AWEL Flow 2.0 construction interface
The flow construction area now features a new AWEL Flow 2.0 interface. Users can manage flows via a new list view (web/pages/construct/flow/index.tsx) that supports creating, copying, and deleting flows, with navigation to either a visual canvas editor or a Python-based 'Libro' editor depending on the flow type. The visual canvas (web/pages/construct/flow/canvas/index.tsx) provides a ReactFlow-based environment for dragging nodes, connecting edges, and saving flow definitions. Additionally, a new App construction page (web/pages/construct/app/index.tsx) has been added to manage application creation, editing, and publishing.
web/pages/construct/flow · high confidence
New AWEL and RAG example workflows
Added several new example scripts in the \examples/awel\ and \examples/rag\ directories to demonstrate DB-GPT's AWEL (Agent Workflow Execution Language) capabilities. These include a data analyst assistant for code fixing and optimization, simple chat DAGs with and without history, LLM client integration for token counting, and various RAG patterns such as schema linking, query rewriting, summarization, and embedding/retrieval workflows. Additionally, dashboard and Olist data loading examples were added to showcase database portability and data visualization setups.
dbgpt-mono · high confidence
New AWEL-based SDK examples for LLM and SQL workflows
Added three new example scripts in the SDK directory demonstrating how to build data processing pipelines using the AWEL (Agent Workflow Execution Language) framework. The examples include a simple LLM interaction, a SQL generation workflow that retrieves database schema and executes queries, and a complex chat data pipeline featuring vector store integration, branching logic, and custom operators.
examples/sdk · high confidence
New Docker base image and build tooling with SQLite and proxy LLM support
The \docker/base\ directory now provides a new Dockerfile and build scripts (\build\_image.sh\, \build\_proxy\_image.sh\) for constructing the application container. The Dockerfile introduces multi-stage builds using \uv\ for dependency management, supports configurable Python versions (default 3.11), and includes options to use Tsinghua mirrors for faster package installation. It installs \sqlite3\ and configures the environment to support SQLite as a local database backend. Additionally, new helper scripts (\run\_sqlite.sh\, \run\_sqlite\_proxyllm.sh\) are provided to easily launch the container with SQLite enabled and, in the case of the proxy script, configured to use external LLM providers via the \proxyllm\ model type.
docker/base · high confidence
New English localization files for chat, knowledge, observability, and flow features
The web interface now includes a comprehensive set of new English translation files (chat.ts, common.ts, flow.ts, knowledgeSource.ts, observability.ts, wiki.ts) that provide user-facing text for the chat interface, knowledge space management (including external source bindings and LLM-Wiki generation), observability dashboards, and flow editor. These additions support the UI for new capabilities such as multi-file upload, MCP connectors, agent tracing, and knowledge base integration.
web/locales/en · high confidence
New Models management page with start/stop controls
A new Models page has been added to the Construct section, allowing users to view a list of models and manage their lifecycle. The interface displays model details such as host, manager host, last heartbeat, health status, and worker type. Users can start, stop, or stop-and-delete models via a dropdown menu on each model card, and create new models using the 'Create Model' button.
web/pages/construct/models · high confidence
New Olist e-commerce example with structured report and evidence flow
Added a new example in \examples/olist\ that demonstrates DB-GPT's report and evidence capabilities using a synthetic Brazilian e-commerce dataset. The entry includes a README explaining the offline synthetic database and instructions for importing real data, a Python script (\build\_demo.py\) to generate deterministic synthetic observations, and a governance document (\olist-analytics-metric-governance.md\) that defines metric definitions, SQL patterns, and analysis rules to ensure report conclusions are bound to SQL and result artifacts rather than model inference.
examples/olist · high confidence
New OpenCode-style agent chat interface with ReAct visualization
The chat component now supports a dedicated 'chat\_agent' mode that renders agent interactions using an OpenCode-style UI. This includes a new container (opencode-agent-chat-container) that integrates with the ReAct Agent API to stream tool execution steps and final answers, and a content renderer (opencode-agent-content) that parses and visualizes these ReAct-style turns, including citations. The system also introduces a new agent-specific content component (agent-content) for rendering standard agent responses with GPTVis and LaTeX support, and updates the main completion flow to handle agent-specific logic and prompt code delivery.
web/components/chat · high confidence
New RAG example scripts for diverse retrieval strategies
The examples/rag directory now includes a comprehensive suite of new demonstration scripts showcasing the expanded RAG capabilities. Users can now explore BM25 full-text retrieval, cross-encoder reranking, and document tree-based retrieval. The examples also cover keyword-based retrieval using Elasticsearch, metadata filtering on vector stores, and database schema RAG. Additionally, new scripts demonstrate GraphRAG with TuGraph integration (including community summaries), query rewriting, summary extraction, and retriever evaluation using metrics like HitRate and MRR. An example for using remote embedding APIs is also provided.
examples/rag · high confidence
New React hooks for chat, agents, connectors, and scheduled tasks
This change introduces a suite of new React hooks in the \web/hooks\ directory to support advanced frontend capabilities. \use-chat.ts\ and \use-react-agent-chat.ts\/\use-react-agent.ts\ handle streaming chat and ReAct agent interactions, including context status monitoring, human-in-the-loop question handling, and SSE parsing. \use-connector-api.ts\ provides CRUD operations and tool fetching for the MCP connector system. \use-scheduled-task.ts\ wraps the scheduled tasks API for task management. \use-subagent-stream.ts\ and its test add support for parsing and restoring parallel sub-agent states and artifacts. \use-summary.ts\ and \use-user.ts\ add document summarization and user info retrieval capabilities.
web/hooks · high confidence
New SQL schema examples for Docker deployments
Added SQL initialization scripts for the \docker/examples/sqls\ directory, providing database setup files for MySQL, SQLite, and Vertica. These include schemas for a student management system, an e-commerce system, and a wide-table order structure, along with corresponding test case definitions and a MySQL user configuration script to enable remote root access.
docker/examples/sqls · high confidence
New Skills management and Conversations list pages in the web UI
The web interface now includes a new Skills management page (web/pages/construct/skills/index.tsx) that allows users to browse, search, filter, and enable/disable skills via the /api/v1/skills endpoints, as well as a new Conversations list page (web/pages/conversations/index.tsx) that displays paginated chat history with search and delete capabilities. Additionally, example code for Excel analysis via Flask has been added to the examples directory.
(repo-wide) · high confidence
New UI components for AWEL flow operators
Added a new Python module \awel\_flow\_ui\_components.py\ in the \examples/awel\ directory that defines example flow operators with rich UI parameter controls. This includes operators for selecting fruits via a search-enabled dropdown, selecting hierarchical addresses via a cascader, and selecting multiple options via checkboxes, allowing users to see how complex UI inputs can be integrated into AWEL flows.
examples/awel · high confidence
New UI components for agent chat content rendering
The chat content area now supports rich rendering of agent interactions, including collapsible thinking processes (VisThinking), structured task lists with status tracking (VisTasks), and multi-tab agent outputs (VisTabs). Users can also view detailed SQL conversion errors with syntax highlighting (VisConvertError), interactive code execution logs (VisCode), and structured data visualizations such as charts, dashboards, and HTML/SVG previews. Additionally, a new ReferencesContent component allows users to inspect source citations for AI responses, and the main chat view now renders custom plugin views with status indicators.
web/components/chat/chat-content · high confidence
New VS Code Dev Container configuration for local development
Developers can now use VS Code's Dev Containers extension to launch a pre-configured development environment based on the eosphorosai/dbgpt-full image. This setup automates the installation of Python 3.11, uv, and development dependencies, configures a zh\_CN.UTF-8 locale, and sets up Oh My Zsh with the powerlevel10k theme and useful plugins. An initialization script handles host user ID/GID mapping to prevent permission issues, and a post-create script ensures the SSH agent and virtual environment are ready for use.
.devcontainer · high confidence
New agent examples and SDK demonstrations
The examples/agents directory now includes a comprehensive suite of runnable demonstrations for the agent SDK. Users can explore single-agent patterns (code, summary, SQL, and ReAct agents), multi-agent orchestration via AutoPlan and AWEL managers, and specialized capabilities like sandboxed code execution, custom tool integration, and MCP/Plugin support. A new \_\main\\_.py entry point allows running all examples sequentially, and documentation for skill-based agents has been added.
examples/agents · high confidence
New app configuration components for AutoPlan, Native App, and Awel Layout modes
Added a suite of new React components in the app construction extra section to support specific team modes. The AutoPlan module (AutoPlan, DetailsCard, ResourcesCardV2, ResourceContentV2) enables users to configure multi-agent workflows with LLM strategies, prompt templates, and dynamic resource bindings. The NativeApp component allows configuration of native application types, models, and parameters. The AwelLayout component provides a workflow selector and preview interface for Awel-based apps. Additionally, RecommendQuestions was added to manage suggested user queries, and config.tsx provides icon mappings for agents and resource types.
web/pages/construct/app/extra · high confidence
New chat header components for model, database, agent, and Excel selection
The chat header now includes dedicated UI components to configure the conversation context: a ModelSelector for choosing LLMs, a DBSelector for picking database parameters, an AgentSelector for switching between AI agents, and a ChatExcel/ExcelUpload module for uploading CSV/XLSX files in the 'chat\_excel' scene. These components are wired into the main Header index, which conditionally renders them based on the current chat scene, allowing users to adjust model, database, agent, or file inputs directly from the chat interface header.
web/components/chat/header · high confidence
New chat page implementation with context management and human-in-the-loop support
The chat interface at web/pages/chat/index.tsx has been replaced with a new implementation that introduces support for agent context management and human-in-the-loop interactions. Users can now engage in conversations where the system may pause to ask clarifying questions, which can be answered or rejected via the new replyQuestion and rejectQuestion controls. The page also integrates dynamic components for database editing and chat containers, manages session history and app parameters (such as temperature and max tokens), and handles routing states for different chat scenes including dashboards and knowledge-based chats.
web/pages/chat · high confidence
New client usage examples for apps, flows, knowledge, and chat
Added seven new Python scripts in the examples/client directory that demonstrate how to use the dbgpt client for common operations. These include CRUD examples for apps, flows, datasources, and knowledge spaces/documents, a chat example showing normal and streaming interactions, and an evaluation example for RAG recall and app answers. Additionally, a new OpenAI-compatible chat client example was added, allowing users to interact with DB-GPT's API v2 endpoint using the standard OpenAI Python library with support for various chat modes and reasoning content display.
examples/client · high confidence
New common UI components for chat, forms, and status display
This change introduces a suite of new reusable components in the common directory to enhance the user interface. Users can now see file processing status via FileStatusIcon, utilize a configurable form system (ConfigurableForm, NestedFormFields) for dynamic parameter inputs, and interact with a new ChatDialog for conversational interfaces. Additionally, the CompletionInput component now supports document uploads and prompt suggestions via PromptBot, while MyEmpty provides consistent empty-state handling with error recovery options.
web/components/common · high confidence
New dashboard chart components for bar, line, pie, and table visualizations
The web application now includes dedicated React components for rendering dashboard charts, located in web/components/chart. This adds support for BarChart, LineChart, PieChart, and TableChart visualizations, all integrated into a unified Chart container that handles layout and type dispatch. These components leverage the G2 charting library for interactive graphics and MUI Joy for structural UI elements, ensuring consistent styling with the application's dark/light theme mode. The implementation includes data processing logic to handle numeric formatting, sorting, and type-value transformations, providing a richer visualization experience for dashboard data.
web/components/chart · high confidence
New database connection management UI components
Added the DatabaseForm and FormDialog React components to the web interface, enabling users to add, edit, and test database connections through a modal dialog. The form supports dynamic parameter fields based on the selected database type, allows adding descriptions, and validates connectivity before saving.
web/components/database · high confidence
New database connection management interface
The Construct page now includes a dedicated interface for managing database connections. Users can view a list of supported database types, add new connections via a modal dialog, edit existing configurations, delete connections, and refresh connection status. The interface displays a count of active connections per database type and handles both supported and unsupported database types, with unsupported options appearing as disabled entries.
web/pages/construct · high confidence
New deployment and maintenance scripts for web builds, model catalogs, and AutoDL environments
Added several new shell and Python scripts to streamline development and deployment workflows. The \build\_web\_static.sh\ script automates the compilation of frontend assets and copies them to the static web directory, handling \.env\ file preservation during the build. \setup\_autodl\_env.sh\ provides a one-command setup for the DB-GPT environment on AutoDL cloud instances, including conda initialization, package installation, and model cloning. \models\_dev\_sync.py\ automates the synchronization of the local model catalog with the upstream models.dev API, merging upstream data with hand-curated provider entries. Additionally, \llama\_cpp\_install.sh\ simplifies the installation of llama-cpp-python by detecting CUDA and CPU instruction set support (AVX/AVX2/AVX512) to select the appropriate binary index, and \run\_llm\_benchmarks.sh\ offers a quick way to execute LLM performance benchmarks with configurable input/output lengths and parallelism.
scripts · high confidence
New flow management UI components for variable, import, export, and save operations
This change introduces a new set of modal components in the flow canvas area to handle core flow lifecycle actions. Users can now add global flow variables with support for string, integer, float, boolean, and reference types via the AddFlowVariableModal. Flow import and export capabilities are provided through ImportFlowModal (uploading JSON/ZIP files) and ExportFlowModal (downloading as JSON or DBGPTS format). Additionally, the SaveFlowModal allows users to save or update flow metadata (name, label, description, editable flag) and control deployment state, while the FlowTemplateModal enables browsing and importing pre-built flow templates.
web/components/flow/canvas-modal · high confidence
New frontend API client layer for chat, evaluation, flows, knowledge, and observability
The web client now includes a comprehensive, structured API client layer that exposes frontend bindings for multiple product areas. This adds dedicated modules for chat (feedback, recommended questions, topic termination), evaluation (dataset and evaluation CRUD, benchmark tasks and results), AWEL flows (CRUD, node management, debugging, import/export, variables), knowledge (document search, chunk editing, recall testing, Git repository sync, knowledge-source bindings, wiki pages/revisions/search), observability (agent health, trace/session queries, metrics, model usage), and prompts (template loading, verification, CRUD). It also centralizes HTTP handling via a shared Axios instance with user-ID headers, long-time API timeouts, and standardized error notifications, while re-exporting all modules from a single index for easier consumption.
web/client · high confidence
New frontend component library for data analysis, charts, and chat UI
The \web/new-components\ directory now provides a new set of React components for the frontend. This includes a \DataAnalyzer\ and \DataPreprocessor\ for statistical analysis and data cleaning, an \AdvancedCharts\ component supporting multiple chart types (line, bar, pie, etc.) with advanced interactions, and a complete chat interface (\ChatPage\, \ChatMessageList\, \ChatHeader\) with auto-scrolling and streaming support. Additionally, new \AppCard\ and \TabContent\ components are introduced for displaying and managing applications.
web/new-components · high confidence
New interactive graph visualization for knowledge spaces
A new page at /construct/knowledge/graph now renders an interactive force-directed graph of knowledge entities using AntV G6. Users can explore nodes and edges with zoom, pan, and drag interactions, while nodes are sized by degree and colored by community. The view highlights connected components via bubble-set overlays and allows hovering to activate related elements, providing a visual summary of the knowledge structure.
web/pages/knowledge · high confidence
New knowledge base management and search interface
The knowledge base UI has been rebuilt with a new sidebar file tree, a dedicated document management panel, and a comprehensive search tools panel. Users can now navigate their knowledge space via a lazy-loaded directory tree (or a fallback document list), manage documents with sync and delete actions, and configure retrieval strategies (embedding, prompt, summary) through a new arguments modal. A new embedded chat component provides streaming responses with citation tracking and a reference panel. Additionally, a recall test modal allows users to validate retrieval configurations, and a Git repository sync form enables direct syncing of code repositories with optional code graph building.
web/components/knowledge · high confidence
New knowledge base management interface with chunk viewing and code graph visualization
This change introduces the core UI pages for the knowledge base module. The main list page (index.tsx) allows users to create, search, and delete knowledge spaces, displaying graph statistics for each. The detail page (detail.tsx) provides a tabbed interface to view document files, browse a wiki index, visualize code graphs via an iframe, and interact with an embedded chat. Additionally, a new chunk list page (chunk/index.tsx) enables users to view, search, and manually add questions to specific document chunks.
web/pages/construct/knowledge · high confidence
New mobile chat interface with feedback and streaming support
A new mobile-optimized chat experience is introduced, featuring a dedicated page layout with a sticky header, message content area, and input controls. Users can now interact via streaming responses (SSE) with the ability to abort ongoing generations. The interface includes a model selector, temperature slider, and resource picker (for files, databases, or knowledge bases) to customize queries. Additionally, users can provide feedback on responses by liking/disliking them, with a drawer to specify reasons for negative feedback, and can copy response text or terminate the current topic.
web/pages/mobile · high confidence
New model management UI components
Added ModelCard and ModelForm components to the web interface. ModelCard displays model status (healthy/unhealthy), host details, and provides a 'Stop Model' action. ModelForm allows users to configure and start new models by selecting worker types (llm, text2vec, reranker), providers, and managing API credentials, with support for preset provider configurations to simplify setup.
web/components/model · high confidence
New models evaluation dashboard and database schema
This release introduces a dedicated models evaluation interface and the underlying database structure to support LLM benchmarking. The new \models\_evaluation\ section in the web UI provides a list of evaluation tasks, a detailed view for benchmark results (including accuracy and execution rates), and a dataset explorer to inspect evaluation data. To support this, the database schema (v0.7.4) adds the \evaluate\_manage\ table for storing benchmark task records and the \benchmark\_summary\ table for storing metric results.
_assets/schema/upgrade/v0\_7\_4, web/pages/models\evaluation · high confidence
New one-line quick installer for DB-GPT
Users can now install DB-GPT with a single command (curl \| bash) that automatically detects the OS, installs the uv package manager, clones the repository, and configures the environment. The installer supports multiple deployment profiles (OpenAI, Kimi, Qwen, MiniMax, GLM, custom, default) to automatically select the correct Python extras and API key environment variables, and allows non-interactive installation via flags like --yes or by providing an existing config file.
scripts/install · high confidence
New parameter input components for flow nodes
The node-renderer module now includes dedicated components for rendering various flow node parameters, including cascader, checkbox, code editor, date picker, input, password, radio, select, slider, textarea, time picker, tree-select, upload, and variables. These components leverage Ant Design UI elements and support dynamic configuration via UI attributes, enabling richer and more interactive parameter inputs within the flow editor.
web/components/flow/node-renderer · high confidence
New prompt construction and management interface
Users can now create, edit, and manage custom prompt templates via a new dedicated section in the Construct workspace. The interface includes a list view for browsing existing prompts with options to edit or delete them, and a detailed editor for adding or modifying prompts, supporting features like variable input, response template configuration, and live LLM output verification.
web/pages/construct/prompt · high confidence
New reproducible dashboard examples and demo data generation tools
Added a new \examples/dashboard\ workspace containing scripts to generate deterministic SQLite databases (including synthetic retail, financial, and research datasets) and pre-built public snapshot fixtures for UI testing. This includes a multi-table Olist e-commerce case with verified gold answers and queries, providing users with ready-to-use, reproducible data sources and dashboard schemas for functional demonstrations and testing.
examples/dashboard · high confidence
New shareable conversation replay page
A new page at /share/\[token\] has been added, allowing users to view shared conversations in a read-only, animated replay format without needing to log in. This page fetches shared session data and visualizes the agent's step-by-step thought, action, and output process, including support for displaying artifacts like code, files, HTML reports, and images generated during the session.
web/pages/share · high confidence
New sidebar navigation and top progress bar components
The application now includes a new side-bar component that provides navigation for Explore, Skills, Datasources, Connectors, and Knowledge sections, along with user controls for toggling light/dark themes and switching between English and Chinese languages. Additionally, a top-progress-bar component has been added to display a loading indicator during route changes and network requests, improving the user's perception of application responsiveness.
web/components/layout · high confidence
New web application shell and construct/observability pages
The web application now uses a new Next.js entry point (\_app.tsx) that wraps all pages with ChatContext, NewTaskProvider, and Ant Design ConfigProvider, enabling unified dark/light theme switching and language support. New pages have been added under /construct for managing agents and scheduled tasks, and under /observability for monitoring agent health, metrics, and traces.
web/pages · high confidence
New web utility modules for agent presentation, flow logic, and model icons
The web/utils area now includes several new modules that enhance the frontend experience: action-display.ts and final-presentation.ts manage how agent thoughts and summary results are streamed and displayed to users; constants.ts provides a comprehensive model icon map and label resolution logic; flow.ts adds utilities for validating and transforming flow node data; react-agent-final.ts and react-sse-parser.ts handle the decoding and parsing of ReAct agent final answers and streaming events; and request.ts centralizes API calls with improved error handling. Corresponding test files (action-display.test.ts, final-presentation.test.ts, react-agent-final.test.ts) have been added to verify these new behaviors.
web/utils · high confidence
Standardized development environment with VS Code Dev Containers and Python 3.11
Developers can now use a standardized, containerized development environment via VS Code Dev Containers, which automatically configures the workspace with Python 3.11, Ruff for formatting and linting, and mypy for type checking. The setup includes pre-commit hooks to enforce code quality and runs tests before commits, while also mounting local models and SSH credentials for seamless integration with the host system.
(repo-wide) · high confidence
Support for multi-file upload and preview in chat sessions
Users can now attach multiple files to a single chat message. The composer displays a new attachment rail with per-file upload progress, error states, and a '+N' overflow collapse. Once uploaded, files appear as read-only cards in the message history, and users can open a dedicated preview panel to view tables, documents, or text content directly within the chat interface.
web/modules/session-files · high confidence
Behavioural changes
2 commits (0 fixes) modifying pilot/examples
A change to existing behaviour in pilot/examples — 2 commits, 2 files: “feat: add dbgpt app start cli (\#2997)”; “feat: Support structured and traceable data analysis reports (\#3295)” (refs \#2997, \#3295).
pilot/examples · medium confidence · unverified
2 commits (1 fix) modifying web/pages/dashboard-share, web/pages/dashboards
A change to existing behaviour in web/pages/dashboard-share, web/pages/dashboards — 2 commits (1 fix), 4 files: “fix(dashboard): unify builder skills and harden template creation (\#3298)”; “feat(dashboard): add dashboard workspace with Next.js 16 toolchain upgrade (\#3278)” (refs \#3298, \#3278).
web/pages/dashboard-share, web/pages/dashboards · medium confidence · unverified
3 commits (0 fixes) modifying pilot/benchmark\_meta\_data
A change to existing behaviour in pilot/benchmark\_meta\_data — 3 commits, 1 file: “feat(benchmark): Support Falcon Text2SQL Datasets LLM benchmark (\#2918)”; “chore: adjust benchmark dataset (\#2940)”; “chore: build web static file” (refs \#2918, \#2940).
_pilot/benchmark\_meta\data · medium confidence · unverified
Database schema for v0.5.1 baseline included in upgrade path
The upgrade script for version 0.5.2 now includes the full SQL definition for the v0.5.1 database schema (in \v0.5.1.sql\). This ensures that the upgrade process has a complete baseline of existing tables—such as \knowledge\_space\, \chat\_history\, and \plugin\_hub\—to correctly apply migrations or verify state, rather than relying on external or missing schema definitions.
_assets/schema/upgrade/v0\_5\2 · medium confidence
Database schema update for v0.5.10 adds domain type to knowledge spaces
The upgrade script for version 0.5.10 modifies the \knowledge\_space\ table by adding a new \domain\_type\ column (varchar(50)) to support categorizing knowledge spaces, such as for financial report analysis. This change is part of the broader schema migration that also includes the full definition of the v0.5.9 database structure, ensuring the database is ready for the new feature requirements.
_assets/schema/upgrade/v0\_5\10 · high confidence
Database schema updated for v0.5.1 with flow error tracking
The upgrade script for version 0.5.1 adds an \error\_message\ column to the \dbgpt\_serve\_flow\ table to store error details. This location also includes the full baseline schema for v0.5.0, establishing the initial database structure including tables for knowledge spaces, documents, chat history, plugins, and prompts.
_assets/schema/upgrade/v0\_5\1 · high confidence
Database schema updates for v0.5.6, v0.5.7, and v0.5.9
This release includes database migration scripts for versions 0.5.6, 0.5.7, and 0.5.9. The v0.5.6 and v0.5.7 upgrades are effectively no-ops for schema changes, with v0.5.6 requiring no action and v0.5.7 only adding a \define\_type\ column to the \dbgpt\_serve\_flow\ table. The v0.5.9 upgrade introduces a new \dbgpt\_cluster\_registry\_instance\ table to support the storage model registry, enabling the registration and management of model instances in a cluster environment with fields for host, port, weight, health checks, and heartbeat tracking.
_assets/schema/upgrade/v0\_5\_6, assets/schema/upgrade/v0\_5\_7, assets/schema/upgrade/v0\_5\9 · high confidence
Database schema updates for v0.6.1 through v0.7.0
This release includes database migration scripts for versions 0.6.1, 0.6.2, and 0.7.0. The v0.6.1 migration adjusts column types and nullability in the \knowledge\_document\, \document\_chunk\, and \chat\_history\_message\ tables. The v0.6.2 migration modifies nullability constraints on \sys\_code\ and \app\_code\ columns in the \gpts\_app\_collection\, \recommend\_question\, and \user\_recent\_apps\ tables. The v0.7.0 migration adds timestamp and extended configuration columns to \connect\_config\, introduces a new \dbgpt\_serve\_model\ table for persisting model worker information, and updates the \user\_code\ column in \recommend\_question\.
(repo-wide) · high confidence
Database schema updates for v0.8.0 beta
The database schema for version 0.8.0 has been updated to support new features including conversation sharing, knowledge management, and plugin integration. The upgrade script introduces a new \share\_links\ table to store conversation share link tokens, and the full schema defines tables for knowledge spaces and documents (\knowledge\_space\, \knowledge\_document\, \document\_chunk\), database connection configurations (\connect\_config\), and a plugin ecosystem (\my\_plugin\, \plugin\_hub\). Additionally, chat history storage has been refactored into separate \chat\_history\ and \chat\_history\_message\ tables to better structure conversation records and details, alongside new tables for user feedback (\chat\_feed\_back\) and prompt management (\prompt\_manage\).
_assets/schema/upgrade/v0\_8\0 · high confidence
Database schema updates for v0.8.1: scheduled tasks, MCP connectors, and message content fix
This release includes database schema changes to support new capabilities and fix existing limitations. It introduces three new tables to enable scheduled chat replay tasks (\dbgpt\_serve\_scheduled\_task\, \dbgpt\_serve\_scheduled\_run\) and persist Model Context Protocol (MCP) connector instances (\connector\_instance\). Additionally, it resolves a data truncation issue by expanding the \content\ column in the \gpts\_messages\ table to \longtext\. The full schema definition for v0.8.1 is also provided for fresh installations.
_assets/schema/upgrade/v0\_8\1 · high confidence
Database schema updates for v0.8.2 introduce code graph analysis and session file tracking
The v0.8.2 database schema adds support for multi-file upload analysis and code graph capabilities. A new \dbgpt\_session\_file\ table enables private persistence for owner-bound session and task files, including metadata like storage URIs and inspection results. The \knowledge\_space\ table gains an \index\_methods\ column to store JSON strings of selected index methods for agentic workflows. Additionally, three new tables—\code\_graph\_vertex\, \code\_graph\_edge\, and \code\_graph\_meta\—are introduced to store AST-extracted code nodes, their structural relationships, and per-knowledge-space graph metadata, enabling code intelligence features.
_assets/schema/upgrade/v0\_8\2 · high confidence
Database schema updates for version 0.6.0
The upgrade path to version 0.6.0 now includes a comprehensive SQL migration script that modifies existing tables and introduces new ones to support enhanced application management, plugin systems, and user interaction tracking. Existing tables such as \chat\_history\, \gpts\_app\, \connect\_config\, and \prompt\_manage\ receive new columns to track app codes, publication status, user identifiers, and response schemas. The migration also creates new tables including \recommend\_question\ for AI-driven suggestions, \user\_recent\_apps\ for tracking usage history, \dbgpt\_serve\_file\ for file storage metadata, \dbgpt\_serve\_variables\ for managing flow and application variables, and dedicated tables for the new plugin hub (\dbgpt\_serve\_dbgpts\_hub\) and user-installed plugins (\dbgpt\_serve\_dbgpts\_my\).
_assets/schema/upgrade/v0\_6\0 · high confidence
Database schema upgrade to v0.7.1 increases column sizes for flow and message data
The upgrade script for version 0.7.1 modifies the database schema to prevent data truncation issues. Specifically, it changes the \action\_report\ column in the \gpts\_messages\ table and the \flow\_data\ column in the \dbgpt\_serve\_flow\ table from \text\ to \longtext\, allowing them to store significantly larger amounts of JSON-formatted data.
_assets/schema/upgrade/v0\_7\1 · high confidence
Database schema upgrade to v0.8.3 with dashboard and tracing support
The metadata database schema is upgraded from v0.8.2 to v0.8.3. This update introduces a new dashboard workspace, adding tables for dashboards, annotations, audit logs, covers, edit versions, folders, live shares, members, operations, and revisions to support collaborative dashboard management. It also adds a \trace\_id\ column to the \chat\_history\_message\ table to enable structured and traceable data analysis. Additionally, the \dbgpt\_serve\_scheduled\_task\ and \dbgpt\_serve\_scheduled\_run\ tables are extended with columns for ownership, resource tracking, leasing, idempotency, and result storage to improve scheduled task reliability and traceability.
_assets/schema/upgrade/v0\_8\3 · high confidence
Introduce AWEL Flow 2.0 frontend components
The flow editor UI has been updated to support the new AWEL 2.0 node model. This change adds a new set of components in the flow area, including a collapsible sidebar (add-nodes-sider) and a popover (add-nodes) for browsing and searching operator and resource nodes, as well as a canvas node (canvas-node) that handles node duplication, deletion, and parameter updates with dependency refreshing. It also introduces a custom edge component (button-edge) that allows users to remove connections directly from the canvas, and a flow card (flow-card) for managing flow lifecycle actions like editing, copying, and deleting. The node parameter handler (node-param-handler) now supports the new AWEL 2.0 UI parameter structure while maintaining backward compatibility with the previous model.
web/components/flow · high confidence
Introduce configurable sandbox execution with local-by-default runtime
DB-GPT now includes a sandbox package that provides a secure, isolated environment for executing agent code (Python, Shell, Node.js, Java, C++, Go, Rust) with resource limits and stateful sessions. The execution backend is configurable via the SANDBOX\_RUNTIME environment variable, supporting Docker, Podman, nerdctl, or local process execution. By default, the sandbox uses local execution; container runtimes are opt-in and will fail closed if the specified backend is unavailable, ensuring that existing workflows continue to run locally without requiring a container setup.
packages/dbgpt-sandbox · high confidence
Unified visual theme and floating button layout fixes
The web interface now uses a unified design system with dedicated light and dark themes, featuring a new color palette (Cobalt primary, Apple success, Ember error) and typography settings defined in the new opencode-theme.css. This change also introduces a z-index hierarchy for floating UI elements to prevent overlapping, with specific responsive adjustments for button sizing and positioning on mobile versus desktop screens.
web/styles · high confidence
Test coverage
Added Playwright E2E tests for dashboard creation, editing, and regression; Added empty test file for Wenxin LLM; Added empty tests package initialization; Added integration tests for OceanBase vector store; Added integration tests for datasource connectors; Added integration tests for graph store adapters and RAG extractors; Added tests for benchmark data manager query timeout handling; Added unit tests for EmbeddingEngine document and URL processing; Added unit tests for FileBasedSkill and SkillLoader; Added unit tests for graph store operations; Added unit tests for pgvector vector store imports; Added unit tests for plugin scanning and access control.
Dependencies
New monorepo structure and dependency manifests for v0.8.2
The project has been restructured into a monorepo using the \uv\ package manager, introducing a root \pyproject.toml\ that defines workspace members including \dbgpt-core\, \dbgpt-client\, \dbgpt-ext\, \dbgpt-serve\, \dbgpt-app\, \dbgpt-sandbox\, and accelerator packages. New \pyproject.toml\ files define dependencies for these packages, specifying Python 3.10+ requirements and optional dependency groups for features like RAG, various database connectors, and GPU acceleration (CUDA/ROCm). The documentation site now uses Docusaurus 3.4.0 with React 18, while the web frontend remains on Next.js 13.4.7 with Ant Design 5.10.0.
(dependencies) · high confidence
Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
How this codebase got here
Score
- CAI 37 → 41 (+3.1)
- Rubric changed (rubric-2026.10.5 → rubric-2026.11.2) — scores are not directly comparable.
Lenses
- Code Health 55 → 55 (+0.0)
- Architecture 59 → 59 (-0.3)
- Maturity 46 → 72 (+25.9)
- Readiness 24 → 28 (+3.6)
- Security 61 → 59 (-2.4)
- Domain Modelling 100 → 100 (+0.0)
- Accessibility 42 → 42 (-0.4)
- Performance 60 → 60 (+0.0)
Resolved (55)
- Change coupling: chunk_db.py ↔ document_db.py (packages/dbgpt-serve/src/dbgpt_serve/rag/models/chunk_db.py)
- Change coupling: controller.py ↔ default_worker.py (packages/dbgpt-core/src/dbgpt/model/cluster/controller/controller.py)
- ClassTooLong: DashboardTemplateSetup (web/new-components/dashboard/DashboardTemplateSetup.tsx)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- DashboardTemplateSetup.DashboardTemplateSetup (cognitive 90) (web/new-components/dashboard/DashboardTemplateSetup.tsx)
- DashboardTemplateSetup.DashboardTemplateSetup (cyclomatic 88) (web/new-components/dashboard/DashboardTemplateSetup.tsx)
- Documentation: written for insiders (docs/data_analysis_planning_agent.md)
- FileTooLong: dashboard/DashboardTemplateSetup.tsx (web/new-components/dashboard/DashboardTemplateSetup.tsx)
- FunctionTooLong: DashboardTemplateSetup.DashboardTemplateSetup (web/new-components/dashboard/DashboardTemplateSetup.tsx)
- High CVE: [GHSA redacted] (web/package-lock.json)
- High CVE: [GHSA redacted] (uv.lock)
- High CVE: [GHSA redacted] (uv.lock)
- High CVE: [GHSA redacted] (uv.lock)
- High CVE: [GHSA redacted] (web/yarn.lock)
- High CVE: [GHSA redacted] (uv.lock)
- …and 35 more
New (133)
- AdvancedCharts.getDashboardAxisConfig (cyclomatic 20) (web/new-components/charts/AdvancedCharts.tsx)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical CVE: [GHSA redacted] (web/package-lock.json)
- Critical CVE: [GHSA redacted] (web/yarn.lock)
- Critical CVE: [GHSA redacted] (web/yarn.lock)
- Critical CVE: [GHSA redacted] (web/yarn.lock)
- Critical CVE: [GHSA redacted] (web/yarn.lock)
- Critical vulnerability: [GHSA redacted] (web/package.json)
- Critical vulnerability: [GHSA redacted] (web/yarn.lock)
- DashboardConfirmationAgent._completion_from_saved_result (cognitive 21) (packages/dbgpt-app/src/dbgpt_app/openapi/api_v1/dashboard/confirmation.py)
- DashboardConfirmationAgent._completion_from_saved_result (cyclomatic 17) (packages/dbgpt-app/src/dbgpt_app/openapi/api_v1/dashboard/confirmation.py)
- FunctionTooLong: DashboardCohortMatrix.DashboardCohortMatrix (web/new-components/dashboard/DashboardCohortMatrix.tsx)
- FunctionTooLong: DashboardTemplatePreview.DashboardTemplatePreview (web/new-components/dashboard/DashboardTemplatePreview.tsx)
- High CVE: [GHSA redacted] (web/yarn.lock)
- …and 113 more
Changes since last survey
- 5 commits — 1 feature/other, 4 fixes
By area
- packages/dbgpt-app — 2 commits
- packages/dbgpt-ext — 1 commit
- packages/dbgpt-serve — 1 commit
- web/new-components — 1 commit
Notable commits
- fix: fix(dashboard): raise agentic completion budget and default generation timeout (#3341)
- fix: fix(dashboard): unify builder skills and harden template creation (#3298)
- fix: fix(knowledge): reject missing graph config (#3352)
- fix: fix: patch path traversal and SQL injection vulnerabilities (#3339)
- change: feat(qdrant): add configurable metadata payload indexes (#3305)
Architecture
- Unchanged — 0 containers · 1 contexts · 0 edges
Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
Survey your own repository
eosphoros-ai/DB-GPT was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.
About this page
- The score is its most recent published measurement, taken on 5 November 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
- Measured at commit 9d50d0a55071a040a8cb7684b3b6acfa13cc7c47 — the exact code this score is about.
- Scored under rubric-2026.11.2 — the same rubric and the same method as every other entry in this index.
- Measured by watchdog.canine.dev using codehealth-analyzer measure-2026.11.2+preprod-f68843d12c58.