VS Code Agent Kanban: Markdown-based task management for AI coding agents
What VS Code Agent Kanban does
VS Code Agent Kanban is a VS Code extension that provides task management specifically designed for developers using AI coding agents like GitHub Copilot. It solves the problem of context rot where planning conversations, decisions, and rationale disappear when chat sessions are cleared or closed.
Key features
- GitOps & team friendly kanban board integration inside VS Code - The entire .agentkanban/ folder is designed to be committed to version control
- Structured plan / todo / implement via @kanban commands - Users type commands like 'plan', 'todo', and 'implement' in the chat window to guide the agent workflow
- Leverages your existing agent harness - Uses your current AI coding agent setup rather than bundling a custom one
- .md task format as permanent source of truth - Each task is a Markdown file with YAML frontmatter that tracks title, kanban lane, and timestamps
How it works
Every task lives in a .agentkanban/tasks/ folder as a .md file. The body contains structured conversation logs using [user] and [agent] markers. Here's the example format from the source:
--- title: Implement OAuth2 lane: doing created: 2026-03-08T10:00:00.000Z updated: 2026-03-08T14:30:00.000Z description: OAuth2 integration for the API --- ## Conversation [user] Let's plan the OAuth2 implementation. We need to support both device code and client credentials flows. [agent] Here's my analysis of the two OAuth2 approaches for your API... [user] Make changes here, here and here [agent] Here's my updated plan ...
The extension creates a clear plan/todo/implement flow where users confirm readiness in GitHub Copilot chat, then the agent harness starts work with all the context preserved in the markdown file.
Benefits for teams
Because tasks are stored as plain text files:
- They're diffable and mergeable naturally - no merge conflicts from opaque binary state
- Teams get shared visibility into what the AI is working on and what's been decided
- Future developers can see not just what was built but why, in the actual words of the planning conversation
- Provides an audit trail for regulated or enterprise environments
This approach is particularly useful for developers who regularly hit context limits with AI agents or need to maintain continuity across multiple work sessions.
📖 Read the full source: HN AI Agents
👀 See Also
Stagent: Open-source ops layer for Claude Agent SDK with local governance and workflow orchestration
Stagent is an open-source, local-first coordination workspace built on top of Claude Agent SDK and Claude API that provides workflow orchestration, budget guardrails, and human-in-the-loop governance for AI agents. It includes 15 product surfaces, 6 workflow patterns, 52+ reusable agent profiles, and runs entirely locally with SQLite.
Two Free Claude Code Skills: Tutorial Generator and Prompt Fixer
Two new free Claude Code skills: create-tutorial generates code reading tutorials from your actual project files, and prompter rewrites typo-filled prompts into actionable instructions. Both are MIT licensed and install via GitHub.
TeamOut AI Agent for Company Retreat Planning
TeamOut has launched an AI agent that plans company events through conversation, handling venue sourcing, vendor coordination, flight cost estimation, itinerary building, and project management. The system uses multiple LLMs and specialized tools to manage planning as a stateful coordination problem.
yburn: Tool to audit and replace unnecessary AI agent cron jobs
yburn is a Python tool that audits AI agent cron jobs and replaces those that don't need LLMs with standalone Python scripts. The creator found 58% of 98 cron jobs were purely mechanical tasks like system health checks and git backups.