$ show esengine/DeepSeek-Reasonix
Reasonix is a local AI coding agent engine written in Go, providing a terminal UI, desktop application (Electron + React), and HTTP server for interacting with large language models (OpenAI, Anthropic, DeepSeek, etc.) to automate software engineering tasks. The kernel manages sessions, context windows, planning, sub-agent delegation, tool execution (file operations, shell, browser, MCP servers), and a sophisticated safety sandbox. It persists conversation history with checkpoints for undo/rewind, tracks token usage, and maintains an evidence ledger for auditing tool effects. The architecture is a monolithic Go binary with multiple frontends: a CLI/TUI, a desktop host that communicates via SSE/JSON, and a REST API server. A plugin system using the Model Context Protocol (MCP) allows extension. Separate Cloudflare Workers handle account management, forum, and crash reporting. The codebase is well-structured with a clean layered architecture (base utilities, domain contracts, state persistence, runtime agent, assembly wiring, frontends), extensive test coverage, and internal linters enforcing quality. It is actively maintained with frequent commits and professional CI/CD pipelines.
The LLM identified 4 files worth highlighting for reuse:
A Go package for static analysis of shell commands. It detects file mutations, heredocs, quote style drift, and pipe tail dropping without executing the command. Demonstrates deep understanding of shell syntax and is safely reusable for any security or auditing tool that needs to inspect shell invocations before execution. Minimal internal dependencies, primarily standard library.
A typed event system with composable sinks (sync, coalesce, diff marking, audit forwarding). Implements a clean observer pattern where components can emit events without knowing consumers. The sink abstractions allow chaining and transformation. Could be extracted as a generic Go event library with minimal dependencies (requires only standard interfaces). Demonstrates non-obvious decoupling design.
A self-contained BM25 scoring implementation with CJK-aware tokenization and snippet generation. Supports field-weighted ranking and is well-parameterized. The algorithm is carefully implemented with proper IDF smoothing and query term contribution. Could be reused in any search or retrieval system with minimal adaptation. Demonstrates strong information retrieval and string handling skills.
Utilities for redacting secrets from API endpoints, configuration maps, and error messages. Provides functions like `RedactEndpoint`, `RedactConfigMap`, and `DiagnosticError` that mask sensitive strings (keys, tokens, passwords) while preserving structure for debugging. Generic enough to be extracted into any application that needs to sanitize logs or user-facing errors. Demonstrates careful security engineering with minimal coupling.