tkmd

tkmd: A Deterministic Program Engine for Task and Project Management

tkmd (with command line interface options as tk or tkmd) is a powerful deterministic program engine designed specifically for developers and AI agents. It organizes and manages tasks using plain Markdown cards with integrated YAML front matter, enabling streamlined project management.

Key Features of tkmd

• Local and Offline Operation: tkmd operates entirely offline, eliminating the need for cloud accounts or subscriptions. It is packaged as a single Rust binary with support for POSIX file locking, ensuring reliability and security.
• Robust AI Agent Management: The system enforces governance by guaranteeing that AI agents cannot arbitrarily mark tasks as complete. All tasks submitted for review undergo automated verification and human sign-off.
• Mathematical Scheduling Algorithms: Utilizing mathematical techniques such as the Critical Path Method (CPM) and Design Structure Matrices (DSM), tkmd computes task dependencies, detects blockers and circular deadlocks, and manages completion schedules efficiently.
• User-Friendly Experience: The tool provides a command-line interface that allows users to easily handle task management through simple commands and shortcuts. For example, transitioning a task to pending close can be achieved through:

 tk review 1 -n "All checks green"  

• Flexible Integration Options: tkmd supports two local setups: integration with Obsidian for Kanban-style task boards or a standalone CLI mode. This flexibility allows users to choose their preferred method of organization.

Task Management Strategies

tkmd enhances task management by implementing structured planning and orchestration mechanisms, which include:

1. Topological Planning: The tool prevents circular dependencies through initial task decomposition and cycle detection in the dependency matrix.
2. Automated Verification Gates: Ensuring that no task can exit the review stage without passing necessary evaluation tests, maintaining accountability for AI agents.
3. Downstream Task Promotion: Automatically managing the state of dependent tasks when blockers are resolved, thus enhancing workflow efficiency.

Privacy and Data Ownership

tkmd prioritizes user privacy, ensuring that all task data is stored in locally accessible Markdown files without external tracking or telemetry. The absence of network requests further guarantees that user data remains completely private and under their control.

Command Overview

tkmd provides a robust command set that simplifies task operations, which can be invoked directly or through efficient shortcuts. Some useful commands include:

• Render Task Dashboard: tk to display the command interface.
• Add Tasks Quickly: Use tk add p1 Fix auth +app @ken due:tomorrow for streamlined task entry.
• Access Task Review: tk review for monitoring pending tasks requiring sign-off.

Conclusion

tkmd stands out as a complete solution for task and project management tailored for both developers and AI agents. Its unique features such as verified workflows, mathematical algorithms, and offline capabilities provide an efficient and trustworthy environment for managing complex development processes.