From experiment runner → product surface.
A pragmatic sequence: stabilize the engine, ship a clean UI for iteration, then expand capabilities (reflection, better evaluation, and multimodal toolchains).
✅ Recently Shipped
Evolution system with comprehensive tooling.
Plugin Registry — Hook-based extensibility with 8 lifecycle hooks for pre/post transformations
Profile Tracking — Automatic agent achievement tracking with specialty detection (10 specialty types)
Reflection Viewer — Browse, search, filter, and export archived reflections with multi-dimensional filters
Codex Promotion — Curated anthology with 6 thematic chapters, cross-references, and auto-categorization
Near-term
Polish and packaging.
• CLI as installable package (pip/conda)
• YAML schema validation for experiments
• Enhanced error messages and logging
• Performance optimizations for large swarms
Mid-term
Advanced evolution features.
• Multi-objective optimization (Pareto fronts)
• Cross-run lineage tracking
• Agent collaboration protocols
• Custom fitness function registry
Product surface
GUI-driven workflow.
• Drag-and-drop experiment builder
• Two-way YAML ↔ Canvas sync
• Monitor tab with “Add to Archive”
• Codex viewer with archetypes + entries
Longer-term
Multimodal + world interaction.
• Vision & video toolchains
• Outcome-constrained learning loops
• Domain-specific swarms (sports, research)
• Stronger governance over memory
How to use this roadmap
Treat each milestone as a test: “Can a new user define an experiment, run it, understand the result, and promote a learned pattern into the Codex?”