Meet the twelve agents
Each specialised agent owns a phase of the lifecycle — triggered automatically, connected to the others, and always working from the same living knowledge.
Twelve specialized agents orchestrate every phase — triggered automatically, connected to each other.
Interviews users, checks capacity, finds duplicates, estimates complexity, suggests priority.
Codebase-aware spec generation with enterprise rules, prior art, and competitor analysis.
Generates visual HTML wireframes using your project's extracted design system.
Generates detailed technical implementation plans with file-level TODOs from approved BUDs.
Detects PR merges, infers status from branches, moves BUD folders, notifies stakeholders.
Aggregates git/PR/bug/chat activity into daily summaries with risk flag detection.
Auto-generates test automation and manual test cases from BUD acceptance criteria and code.
Links bugs to BUDs via vector search, monitors thresholds, triggers reassignment.
Reassigns devs to bug review, rotates QA to next BUD, rebalances workloads.
Cycle time analysis, estimate vs actual comparison, pattern matching, retrospective generation.
Assigns the best-fit developer based on skill scores, capacity, and module expertise.
Rebuilds skill profiles from git/BUD/bug history, scores 0–1.0, detects bus factor risks.