A lightweight, repo-centric development tool for orchestrating AI coding agents, Sudocode addresses the challenges of managing context, coordination, and long-horizon tasks in human-AI collaborative software development. By integrating directly into your repository, Sudocode enables agents to track context, collaborate on complex tasks, and allows developers to direct agent work more effectively, reducing the need for constant oversight.

How It Works

Sudocode employs a git-native agent memory system, distinguishing between "Specs" (human-defined requirements and objectives) and "Issues" (agent runtime context and implementation plans). Both are version-controlled within your git repository, making them human and agent-editable. This approach allows agents to autonomously manage their context, leave "breadcrumbs" for other agents, and coordinate efforts. Specs and Issues form a structured, executable graph, enabling advanced agent coordination and automation by defining what needs to be done and how agents plan to implement it.

Quick Start & Requirements

• Installation: npm install -g sudocode or use npx sudocode.
• Initialization: Run sudocode init in your project directory.
• Local Server: Start with sudocode server.
• Prerequisites: Node.js environment (implied by npm/npx).
• Documentation: Refer to the official Quick Start Guide for tutorials.

Highlighted Details

• Context-as-Code: Treats your git repository as a distributed context database, versioning user intent (Specs), agent tasks (Issues), and execution trajectories.
• Dual Representation: Entities (Specs, Issues) exist in both human-readable Markdown+YAML and machine-optimized JSONL+SQLite formats, with automatic synchronization and git-friendly storage.
• Graph-Based Planning: Enables topological ordering for execution, hierarchical task management, and concurrent agent operations based on defined dependencies.
• Bidirectional Linking: Establishes and tracks relationships (e.g., blocks, related, parent-child) between Specs and Issues in both directions for comprehensive traceability.
• Agent Feedback Loops: Allows agents to provide feedback on Specs during execution, bridging implementation details back to high-level requirements.
• IDE Integration: Seamlessly opens agent execution worktrees directly in popular IDEs like VS Code, Cursor, and IntelliJ for streamlined code review and modification.

Maintenance & Community

No specific details regarding maintainers, community channels (like Discord/Slack), or project sponsorships were found in the provided README excerpt.

Licensing & Compatibility

The license type and any compatibility notes for commercial use or closed-source linking are not specified in the provided README excerpt.

Limitations & Caveats

Integration with external issue trackers like Jira and Linear is listed as "Coming soon," indicating ongoing development and feature expansion. The system relies heavily on Git for distribution and conflict resolution, which may introduce complexities depending on team workflows.