Roadmap

This roadmap outlines the evolution of Zene from a high-performance engine into a production-grade AI coding platform.

1. Advanced Knowledge & Context

• LSP Integration: Connect to Language Servers for "Go to Definition" and real-time diagnostics before running code.
• Context Graph: Build a cross-file knowledge graph to automatically include relevant model/schema definitions during edits.
• Context Compaction: Implement smart summarization of long session histories to maintain focus.

2. Interactive Human-in-the-Loop (HITL)

• Proactive Clarification: Agent pauses and asks for hints when stuck on a repetitive error.
• Safety Gateways: Explicit confirmation for high-risk actions (rm -rf, git push).
• Web Dashboard: An interactive UI to review diffs and guide the agent mid-task.

3. Toolchain & Environment

• Multi-language Sandboxing: Extend Python-style isolation to Node.js and Rust environments.
• Docker Tooling: Ephemeral containers for complex services (DBs, Redis) during integration testing.
• AST-based Editing: Move beyond text-based patches to precise AST manipulation for zero-conflict edits.

4. Observability & Developer Experience

• TUI Dashboard: A rich terminal UI showing task progress, DAG status, and real-time logs.
• Artifact Summaries: Automated reports on latency, token costs, and generated assets per session.
• Integrated Benchmarking: Automated evaluation suites to measure Agent success rates on real-world coding tasks.