AI coding experiment with Deno for code generation
Top 49.3% on sourcepulse
This project serves as an experimental playground for combining AI coding agents with Deno, aiming to improve code generation quality and efficiency by defining coding rules and modes for Deno projects. It targets developers and AI researchers interested in structured, AI-assisted code development within the Deno ecosystem.
How It Works
The core of the project lies in defining coding rules and implementation modes in Markdown files within the .cline
directory. A build.ts
script then processes these files to generate .clinerules
and .roomodes
(JSON) files. These generated files act as configuration for AI coding agents, guiding them to produce type-safe, Deno-idiomatic code according to specified best practices and implementation strategies like script mode, module mode, or test-first mode.
Quick Start & Requirements
git clone <repo_url>
deno cache --reload deps.ts
deno run --allow-read --allow-write .cline/build.ts
Highlighted Details
type-predictor
module: Analyzes JSON data to predict types and generate Zod schemas for runtime validation.zodcli
module: Provides type-safe command-line argument parsing using Zod, with automatic help generation and subcommand support.ts-callgraph
tool: Generates call graphs for TypeScript code.Maintenance & Community
The project appears to be primarily maintained by mizchi
. No specific community links (Discord, Slack) or roadmap are explicitly mentioned in the README.
Licensing & Compatibility
Limitations & Caveats
The project is experimental, with several components listed as "in development" or "initial stage" (e.g., type-predictor
, memorybank
, documentation, CI/CD). Full documentation and test coverage are stated as future goals.
4 months ago
Inactive