agentcode-cli 1.0.0__py3-none-any.whl

agentcode_cli-1.0.0.dist-info/METADATA

@@ -0,0 +1,303 @@
+Metadata-Version: 2.4
+Name: agentcode-cli
+Version: 1.0.0
+Summary: An open, multi-model agentic coding CLI — inspired by Claude Code
+Author: Vignesh Pai
+License: MIT License
+        
+        Copyright (c) 2026 Vignesh Pai
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+        The above copyright notice and this permission notice shall be included in all
+        copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+        
+Project-URL: Homepage, https://github.com/vigp17/AgentCode
+Project-URL: Repository, https://github.com/vigp17/AgentCode
+Project-URL: Bug Tracker, https://github.com/vigp17/AgentCode/issues
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Environment :: Console
+Classifier: Topic :: Software Development
+Classifier: Topic :: Utilities
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: litellm>=1.40.0
+Requires-Dist: rich>=13.0.0
+Requires-Dist: python-dotenv>=1.0.0
+Requires-Dist: mcp>=1.0.0
+Provides-Extra: dev
+Requires-Dist: build; extra == "dev"
+Requires-Dist: twine; extra == "dev"
+Dynamic: license-file
+
+# AgentCode
+
+An open, multi-model agentic coding CLI — inspired by Claude Code.
+
+## Architecture
+
+```
+┌─────────────────────────────────────────────────┐
+│                  cli.py (UI)                    │
+│  REPL loop · slash commands · Rich terminal UI  │
+│  AGENTCODE.md · session persistence             │
+└──────────────────────┬──────────────────────────┘
+                       │
+┌──────────────────────▼──────────────────────────┐
+│               agent.py (Brain)                  │
+│  Agentic loop · context management · permissions│
+│  Hooks · subagents · LLM-powered compaction     │
+│                                                 │
+│   while needs_follow_up:                        │
+│     1. Send messages + tools → LLM              │
+│     2. If tool_calls → execute, append, loop    │
+│     3. If text only  → done                     │
+│                                                 │
+│   LiteLLM ──→ Claude / GPT / Gemini             │
+└──────────────────────┬──────────────────────────┘
+                       │
+          ┌────────────┴────────────┐
+          │                         │
+┌─────────▼───────────┐   ┌─────────▼───────────┐
+│    tools.py (Hands) │   │  mcp_client.py       │
+│  read_file          │   │  Connect to MCP      │
+│  write_file         │   │  servers and expose  │
+│  edit_file          │   │  their tools to the  │
+│  run_command        │   │  agent loop.         │
+│  list_directory     │   └─────────────────────┘
+│  search_files       │
+│  search_text        │
+│  git_status/diff    │
+│  git_log/commit     │
+│  git_branch/push    │
+│  spawn_subagents    │
+└─────────────────────┘
+```
+
+## Quick Start
+
+```bash
+# Option A: Install from PyPI
+pip install agentcode-cli
+
+# Option B: Install from source
+git clone https://github.com/vigp17/AgentCode.git
+cd AgentCode
+pip install -r requirements.txt
+```
+
+```bash
+# Add your API keys to a .env file in your project directory
+ANTHROPIC_API_KEY=sk-ant-...
+OPENAI_API_KEY=sk-...
+GEMINI_API_KEY=...
+```
+
+```bash
+# Run
+agentcode                               # Interactive REPL
+agentcode "fix the failing tests"       # One-shot mode
+agentcode --model gpt-4o               # Use a different model
+agentcode --init-settings              # Create a settings file
+```
+
+## Supported Models
+
+Via [LiteLLM](https://docs.litellm.ai/docs/providers), AgentCode works with any model that supports function calling — including Claude, GPT, Gemini, and select open-source models via Ollama.
+
+| Provider   | Model String                    | API Key Env Var         |
+|------------|---------------------------------|-------------------------|
+| Anthropic  | `claude-sonnet-4-6` (default)   | `ANTHROPIC_API_KEY`     |
+| OpenAI     | `gpt-4o`                        | `OPENAI_API_KEY`        |
+| Google     | `gemini/gemini-2.5-pro`         | `GEMINI_API_KEY`        |
+
+## Cost-Aware Routing
+
+AgentCode automatically picks the cheapest model that can handle the task:
+
+| Tier   | Anthropic          | OpenAI       | Gemini                  |
+|--------|--------------------|--------------|-------------------------|
+| Light  | Haiku 4.5          | GPT-4o Mini  | Gemini 2.0 Flash        |
+| Medium | Sonnet 4.6         | GPT-4o       | Gemini 2.5 Flash        |
+| Heavy  | Opus 4.6           | GPT-5.5      | Gemini 2.5 Pro          |
+
+Simple questions go to cheap/fast models. Complex multi-file tasks go to powerful ones. Use `--no-route` to always use the specified model.
+
+## Tools
+
+### File & Shell
+
+| Tool             | Description                          | Permission |
+|------------------|--------------------------------------|------------|
+| `read_file`      | Read file contents with line numbers | Auto       |
+| `write_file`     | Create or overwrite a file           | Ask        |
+| `edit_file`      | Surgical find-and-replace edit       | Ask        |
+| `run_command`    | Execute a bash command               | Ask        |
+| `list_directory` | Tree view of directory structure     | Auto       |
+| `search_files`   | Find files by glob pattern           | Auto       |
+| `search_text`    | Grep for text across files           | Auto       |
+
+### Git
+
+| Tool          | Description                              | Permission |
+|---------------|------------------------------------------|------------|
+| `git_status`  | Show working tree status                 | Auto       |
+| `git_diff`    | Show staged or unstaged changes          | Auto       |
+| `git_log`     | Show recent commit history               | Auto       |
+| `git_commit`  | Stage files and create a commit          | Ask        |
+| `git_branch`  | List, create, or switch branches         | Ask        |
+| `git_push`    | Push commits to a remote                 | Ask        |
+
+### Subagents
+
+| Tool              | Description                                      | Permission |
+|-------------------|--------------------------------------------------|------------|
+| `spawn_subagents` | Run multiple agents in parallel on subtasks      | Auto       |
+
+**Permission model:** Read-only tools auto-approve. Write/execute tools ask before running (unless `--auto-approve` / `-y` flag is set).
+
+## Slash Commands
+
+| Command                  | Description                                      |
+|--------------------------|--------------------------------------------------|
+| `/model <name>`          | Switch LLM model on the fly (disables routing)   |
+| `/route`                 | Show or toggle cost-aware routing                |
+| `/cost`                  | Show session cost breakdown                      |
+| `/mcp`                   | Manage MCP server connections                    |
+| `/mcp list`              | Show connected servers and tool counts           |
+| `/mcp add <server>`      | Connect a server (prompts for credentials)       |
+| `/mcp remove <server>`   | Disconnect a server                              |
+| `/clear`                 | Reset conversation and delete saved session      |
+| `/compact`               | Force LLM-powered context compaction             |
+| `/tokens`                | Show estimated token usage                       |
+| `/init`                  | Create an AGENTCODE.md template                  |
+| `/settings`              | Show resolved settings and active config files   |
+| `/help`                  | Show help                                        |
+| `/exit`                  | Quit                                             |
+
+## Session Persistence
+
+Conversations are automatically saved to `.agentcode_session.json` in your project directory and resumed on next launch. Use `/clear` to start fresh.
+
+## Hooks
+
+Run shell commands before or after any tool call. Create `.agentcode/hooks.json`:
+
+```json
+{
+  "post_edit_file": "prettier --write \"$AGENTCODE_PATH\"",
+  "post_write_file": "prettier --write \"$AGENTCODE_PATH\"",
+  "pre_run_command": "echo \"Running: $AGENTCODE_COMMAND\""
+}
+```
+
+Supported keys: `pre_<toolname>`, `post_<toolname>`, `pre_tool` / `post_tool` (wildcard). Tool args are passed as `AGENTCODE_<ARG>=value` env vars. Global hooks go in `~/.agentcode/hooks.json`.
+
+## MCP Support
+
+Connect to any [MCP server](https://modelcontextprotocol.io) using the `/mcp add` command — no manual config editing needed:
+
+```
+/mcp add github       # prompts for GitHub token, connects immediately
+/mcp add filesystem   # connects to current directory, no credentials needed
+/mcp add postgres     # prompts for connection string
+/mcp add sqlite       # prompts for database file path
+/mcp list             # show connected servers and tool counts
+/mcp remove github    # disconnect a server
+```
+
+Servers connect live without restarting. Config is saved to `.agentcode/mcp.json` and reloaded automatically on next launch.
+
+MCP tools are exposed as `mcp__<server>__<toolname>` (e.g. `mcp__github__create_issue`) and available to the LLM alongside built-in tools. Connected servers are shown in the banner on startup.
+
+**Advanced:** You can also edit `.agentcode/mcp.json` directly for custom servers:
+
+```json
+{
+  "mcpServers": {
+    "my-server": {
+      "command": "npx",
+      "args": ["-y", "@myorg/mcp-server"],
+      "env": {"API_KEY": "..."}
+    }
+  }
+}
+```
+
+Global config goes in `~/.agentcode/mcp.json`.
+
+## Subagents
+
+AgentCode can spawn parallel agents to work on independent subtasks simultaneously. Just ask naturally:
+
+> "Analyze agent.py, router.py, and tools.py in parallel and summarize each one"
+
+The agent calls `spawn_subagents` internally, runs up to 5 agents in parallel, and returns combined results.
+
+## AGENTCODE.md
+
+AgentCode loads project-level instructions from `AGENTCODE.md` in your project directory (and global config from `~/.agentcode/AGENTCODE.md`), injecting them into the system prompt automatically.
+
+Run `/init` to generate a starter template, then edit it to define your project's coding standards, preferences, and constraints.
+
+## How to Extend
+
+### Add a new tool
+
+1. Add the function schema to `TOOL_DEFINITIONS` in `tools.py`
+2. Implement the function (e.g., `_my_tool(...)`)
+3. Register it in `TOOL_MAP`
+4. If it should require user approval, omit it from `permissions.auto_approve` in `.agentcode/settings.json`
+
+## Environment Variables
+
+| Variable                   | Description                       | Default             |
+|----------------------------|-----------------------------------|---------------------|
+| `AGENTCODE_MODEL`          | Default model                     | `claude-sonnet-4-6` |
+| `AGENTCODE_MAX_ITERATIONS` | Max tool-call iterations per turn | `25`                |
+| `ANTHROPIC_API_KEY`        | Anthropic API key                 | —                   |
+| `OPENAI_API_KEY`           | OpenAI API key                    | —                   |
+| `GEMINI_API_KEY`           | Google Gemini API key             | —                   |
+
+## Publishing to PyPI
+
+```bash
+# 1. Install build tools
+pip install build twine
+
+# 2. Bump the version in pyproject.toml, then build
+python -m build
+
+# 3. Upload to PyPI
+twine upload dist/*
+```
+
+You'll be prompted for your PyPI credentials on first upload. Use an API token from [pypi.org/manage/account](https://pypi.org/manage/account) for security.
+
+For test uploads before going live:
+
+```bash
+twine upload --repository testpypi dist/*
+pip install --index-url https://test.pypi.org/simple/ agentcode
+```
+
+## License
+
+MIT

agentcode_cli-1.0.0.dist-info/RECORD

@@ -0,0 +1,12 @@
+agent.py,sha256=zcYfTAarYhj1hdLc9DH5sWs6Qqqp1VpmrHGJfHSJZIQ,17118
+cli.py,sha256=naUTUUgvL2DJRTTZzxQCOi6bjZ7rmtmpwlqaGDQ4eQk,24677
+mcp_client.py,sha256=2PviTqJtXM4UC_fsYLbAOAfWJvayWy7Q8VOQIIsDiqQ,6710
+router.py,sha256=kjPpF1js6mA7FN2x_NleyN-BLeVzRAy-Q0wbhjelQzM,11109
+settings.py,sha256=Qjc3tiVbT1cqIrnQW6m2UG8Xsvqsxl9qXPMTueqwn50,6903
+tools.py,sha256=MBYy0OeSIZjqyGOezaMvZ6AMW3WRDESU-2u48sifd0Q,24741
+agentcode_cli-1.0.0.dist-info/licenses/LICENSE,sha256=BqTzyKKaSaVQoumXzhYCj1UgOSPCgvn-sxV6BIuT558,1068
+agentcode_cli-1.0.0.dist-info/METADATA,sha256=gJjlU0B67K1iMGnSgYVk901xeCuaHc12RSXL5Cs6WM0,13389
+agentcode_cli-1.0.0.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
+agentcode_cli-1.0.0.dist-info/entry_points.txt,sha256=xP_zeySufuVhL5v10_EqCooKRSxGYB8QuRVNUj4_m1E,39
+agentcode_cli-1.0.0.dist-info/top_level.txt,sha256=5jYoRO_eF5cWLQAvnel4HxTjCgsG0iBHpOkcTn3Gqv8,43
+agentcode_cli-1.0.0.dist-info/RECORD,,

agentcode_cli-1.0.0.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (82.0.1)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

agentcode_cli-1.0.0.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+agentcode = cli:main

agentcode_cli-1.0.0.dist-info/licenses/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Vignesh Pai
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

agentcode_cli-1.0.0.dist-info/top_level.txt

@@ -0,0 +1,6 @@
+agent
+cli
+mcp_client
+router
+settings
+tools