@anastaha/codeforge
v1.1.0
Published
AI coding agent CLI — reads, edits, patches and runs commands in your repo via LLM
Maintainers
Readme
@anastaha/codeforge
AI coding agent that runs in your terminal. It reads, edits, and runs commands in your repo using LLMs.
Install
npm install -g @anastaha/codeforgeSetup
- Get a free API key from OpenRouter
- Set the key:
export LLM_API_KEY=your-key-here- Run in any git repo:
codeforgeWhat it can do
- Read & Edit files with surgical precision
- Search your codebase (glob patterns, regex grep)
- Run shell commands (git, npm, tests, builds)
- Search the web for docs and solutions
- Delegate tasks to sub-agents that work in parallel
- Stream responses in real-time
Tools
| Tool | Description |
|------|-------------|
| Read | Read file contents |
| Edit | Find & replace exact strings in files |
| Write | Create or overwrite files |
| Glob | Find files by pattern |
| Grep | Search file contents by regex |
| ListDirectory | List directory contents |
| Bash | Run shell commands |
| WebSearch | Search the web |
| Delegate | Spawn a sub-agent for parallel tasks |
Commands
| Command | Description |
|---------|-------------|
| /help | Show commands |
| /model <name> | Switch LLM model |
| /reset | Reset conversation |
| /compact | Compact message history |
| /clear | Clear screen |
| /quit | Exit |
Configuration
Set via environment variables or .env file:
| Variable | Default | Description |
|----------|---------|-------------|
| LLM_API_KEY | — | Required. OpenRouter API key |
| CODEFORGE_MODEL | nvidia/nemotron-3-nano-30b-a3b:free | LLM model to use |
| CODEFORGE_TIMEOUT | 120 | API timeout (seconds) |
| CODEFORGE_BASH_TIMEOUT | 120 | Shell command timeout (seconds) |
| CODEFORGE_MAX_ROUNDS | 25 | Max tool call rounds per query |
| LANGSEARCH_API_KEY | — | Optional. For web search |
License
MIT
