1C AI Development Environment
A 1C development environment with a 10-tool MCP server: 11 BSL analyzers, JSON DSL to XML compilers, a dependency graph and EPF builds with no 1C installed
High risk
We rate an entry high when the tool writes to external systems, handles money, production databases or secrets, or runs arbitrary commands. The CLI installs it only with your consent.
Why this level
- build_epf and the code generators create and change configuration and extension files on disk
- cfe_patch_method generates intercepting methods in an extension, changing how the configuration behaves
Install
Manual install
git clone https://github.com/Pradushkoai/1c-ai-dev-env.git && cd 1c-ai-dev-env && pip install -e ".[dev,mcp]"Optionally follow with bash install.sh to install the BSL Language Server.
This is third-party code. Review the repository files before installing.
What it does
The project parses a 1C configuration XML export and builds five indexes: metadata (35 object types), a reference of platform BSL methods, SKD schemas, form structure, and a networkx dependency graph. On top of the indexes run 11 BSL analyzers with over 150 rules (security, coding standards, transactions, queries, complexity metrics, architecture), five JSON DSL to XML compilers for metadata, forms, SKD, templates and roles, and extension operations (borrowing an object, generating interceptor methods, diffing an extension). Separately, EpfFactory builds an external .epf processing from scratch with no 1C running, via templates and v8unpack. All of this is available through a 19-command CLI and the project's own MCP server with 10 tools for Cursor, Claude Desktop, VS Code and JetBrains. The project is at Beta status, with each subsystem explicitly marked for stability, and 1595 passing tests.
Who it is for. For 1C developers who want a full analytical and code-generation environment around an agent, not just a reference or search tool.
Good fit when
- You need a security and BSL coding-standards audit across dozens of rules at once
- You need to generate metadata, forms, SKD or roles from a JSON description without hand-editing XML
- You need a metadata dependency graph to find dead code or cycles
Not a fit when
- You need only stable minimal functionality: some subsystems (EDT parser, RAG with Ollama, SaaS/Enterprise) are marked experimental or frozen
- You lack Python 3.10+ and, for the full analyzer set, Java 17+ for the BSL Language Server
- You need to work against a live 1C base, not a configuration XML export
Example request
Check the ОбщегоНазначения.bsl module with all seven analyzers and show only new errors compared to the baselineLimitations
The project itself states Beta status, not production-ready, with an explicit per-subsystem stability matrix: EpfFactory, the SARIF reporter, OpenSpec and the Session manager are marked Beta, the EDT parser and RAG with Ollama are marked Experimental, and SaaS/Enterprise/Plugin are frozen and moved to experimental/. Building an EPF uses a v8unpack workaround rather than an official API. The full set of 150+ analysis rules requires the Java-based BSL Language Server installed.
How to disable. Remove the 1c-ai server from your client's mcp.json and run docker compose down if Docker was used.
MCP
- Transport
- stdio
- Authentication
- not required
Security check
- build_epf and the code generators create and change configuration and extension files on disk
- cfe_patch_method generates intercepting methods in an extension, changing how the configuration behaves
README in short
The README lists 11 BSL analyzers with their rule counts, metadata parsing components, five DSL compilers, CFE extension operations, infrastructure tools (SARIF, OpenSpec, Session manager, EpfFactory), an explicit Stability Matrix per subsystem, installation via pip or Docker, the full list of 19 CLI commands by category, MCP setup for Cursor, Claude Desktop, VS Code and JetBrains with a 13-category tool table, and the project's directory architecture. It separately instructs an AI agent to read AGENTS.md before working.
FAQ
Can it be used without the BSL Language Server?
Yes, install.sh for it is optional, but without it the analyze_bsl check with 187 diagnostics is unavailable.
How do you know which subsystems are safe to use in production?
By the Stability Matrix table in the README: Stable subsystems are tested, Beta and Experimental ones need caution.
Related
A skills library that gives coding agents a development process: brainstorming, planning, TDD, subagents and code review
Skills for real engineers by Matt Pocock
Skills For Real Engineers
Small composable skills for engineering with agents: plan grilling, TDD, bug diagnosis, code review and architecture
GitHub toolkit for spec-driven development: the specify CLI adds agent commands and skills to a project, from principles to implementation
Reference MCP servers
Model Context Protocol servers
Official reference MCP servers: Filesystem, Fetch, Git, Memory, Sequential Thinking, Time and Everything