name: Feature Request
about: Suggest an idea for this project
title: '[FEATURE] Add --watch flag for hot-reloading agent during development'
labels: enhancement
assignees: ''
Feature Description
Add a --watch / -w flag that automatically reloads the agent when the agent file or its dependencies are modified, similar to nodemon for Node.js development.
Problem/Motivation
Is your feature request related to a problem?
When developing and iterating on agents, I have to manually exit the chat loop and restart it every time I make a code change. This slows down the development cycle significantly.
Example: "I'm frustrated when I need to test changes to my agent because I have to exit the chat, restart chat_loop, and lose my conversation context every time."
Proposed Solution
Add a watch mode that monitors the agent file and automatically reloads when changes are detected.
Example:
# Start in watch mode
chat_loop myagent --watch
# Or with short flag
chat_loop path/to/agent.py -w
# Output when changes detected:
# 🔄 Agent file modified, reloading...
# ✓ Agent reloaded successfully
Alternatives Considered
- Manual restart - Current approach, too slow
- External watcher script - Could use entr/watchexec but requires additional setup
- Editor integration - Would require per-editor plugins
Benefits
Who would benefit from this feature and how?
- Agent developers: Significantly faster iteration cycle during development
- QA testers: Can test changes immediately without losing context
- Anyone prototyping: Enables rapid experimentation with agent behavior
Implementation Ideas
- Use
watchdog library to monitor file changes
- Watch the agent file itself plus any relative imports
- Option to watch entire directory:
--watch-dir path/
- Debounce rapid changes (wait 500ms after last change)
- Preserve conversation history across reloads (optional)
- Show clear indication when reload happens
Suggested implementation approach:
from watchdog.observers import Observer
from watchdog.events import FileSystemEventHandler
class AgentReloader(FileSystemEventHandler):
def on_modified(self, event):
if event.src_path == agent_path:
reload_agent()Additional Context
Similar to:
nodemon for Node.jscargo watch for Rust- Django's auto-reload
- Flask's debug mode
This is a common pattern in modern development tools and would align well with the project's focus on developer experience.
Priority
How important is this feature to you?
name: Feature Request
about: Suggest an idea for this project
title: '[FEATURE] Add --watch flag for hot-reloading agent during development'
labels: enhancement
assignees: ''
Feature Description
Add a
--watch/-wflag that automatically reloads the agent when the agent file or its dependencies are modified, similar to nodemon for Node.js development.Problem/Motivation
Is your feature request related to a problem?
When developing and iterating on agents, I have to manually exit the chat loop and restart it every time I make a code change. This slows down the development cycle significantly.
Example: "I'm frustrated when I need to test changes to my agent because I have to exit the chat, restart
chat_loop, and lose my conversation context every time."Proposed Solution
Add a watch mode that monitors the agent file and automatically reloads when changes are detected.
Example:
Alternatives Considered
Benefits
Who would benefit from this feature and how?
Implementation Ideas
watchdoglibrary to monitor file changes--watch-dir path/Suggested implementation approach:
Additional Context
Similar to:
nodemonfor Node.jscargo watchfor RustThis is a common pattern in modern development tools and would align well with the project's focus on developer experience.
Priority
How important is this feature to you?