See it

The platform

The Living Tree, the BUD board, estimation, quality loops, knowledge, and the gamification layer — the surfaces where Agent-Driven Development happens.

See it

The Living Tree, the BUD board where features move through their lifecycle, and the Feature registry of what shipped.

Living Tree dashboard — 3D visualisation of your repos

Feature registry — BUDs that graduated to shipped features

AI-PERT + Monte Carlo Estimation

Story pointsPlanning poker — replaced by probabilistic, per-phase delivery predictions that improve automatically.

AI generates optimistic, likely, and pessimistic estimates (PERT) for each phase. 10,000 Monte Carlo simulations produce P50/P70/P85 confidence dates. Developer skill profiles, backlog depth, and workload are factored in.

Example Prediction

Feature: Notification Redesign (Complexity: 3/5)

Developer: Alice (backend: 0.92, frontend: 0.35)

Go-live: 70% by Apr 25 | 85% by May 2

PERT Analysis

10K Simulations

Skill-Aware

BUD (Business Understanding Document)

Every feature lives in one BUD — a Business Understanding Document holding the spec, tech spec, test plan, acceptance criteria, and full history. Replaces scattered Jira tickets, Google Docs, and Notion pages.

bud

design

tech_arch

development

testing

uat

prod

closed

Contains spec, tech spec, test plan, acceptance criteria, and metadata

Any stage can return to BUD (e.g., post-deployment bugs)

Bug classification on reopen: "missed feature" vs "development bug"

Full history tracked: stage transitions, assignees, reopens, bugs

Vector-indexed for semantic search by all agents

Smart Quality Loops

Auto-healing bug management that prevents quality debt from accumulating.

Bug Threshold

complexity × multiplier — configurable per org. When exceeded, auto-reassignment triggers.

Auto-Reassignment

Original dev moves to bug review, QA moves to next waiting BUD.

Feature Reopening

External bugs reopen the Feature and restart the flow from triage.

Auto-Classification

Each bug classified as "missed feature" vs "development bug" — drives different fix paths.

Knowledge Capture

Every bug fix adds to the knowledge base — prevents the same bug class from recurring.

Backlog Intelligence

Smart backlog management driven by data, not gut feelings.

Capacity-Aware Triage

Triage Agent deprioritizes or defers items based on real-time team capacity.

Dynamic Reassignment

Reassignment Agent shuffles work based on shifting business demand.

Customer Priority Scoring

ARR + severity + tier drives backlog ordering automatically.

Best-Fit Developer

Skill Agent recommends the best-fit developer for each backlog item.

Real-Time Utilization

Per-developer capacity tracking ensures balanced workloads.

Knowledge That Grows

A 4-layer knowledge architecture that replaces stale wikis with living, auto-synced knowledge.

Git Repos

Source code + per-repo CLAUDE.md (syncs every 15 min)

Agent Skills

Org standards, design guidelines, API patterns (syncs on change)

Central DB

BUDs, enterprise rules, architecture decisions (real-time)

Vector Search

Semantic search across everything (auto-indexed)

Why This Beats Confluence

Auto-synced from source — not manually maintained

Semantically searchable — not keyword search

Always current — daily staleness detection

Integrated into agent prompts — agents always have latest context

Dev Skill Maintenance

The Skill Agent rebuilds developer profiles daily — no manual updates, no stale resumes.

Daily Profile Rebuilds

Analyzes git history, BUD assignments, and bug fixes to build skill scores (0–1.0) per module.

Bus Factor Alerts

Detects modules touched by only one person — flags knowledge concentration risk.

Assignment Recommendations

Recommends developers for new BUDs based on expertise match + available capacity.

Evolving Skills

Skills grow automatically as developers contribute — no manual profile updates needed.

The gamification layer

The Skill Agent rebuilds developer profiles nightly. Skills compound, badges unlock, and the leaderboard reflects what people shipped — not how many tickets they touched.

Unlocks — badges earned through contribution patterns

The Living Tree: your org as a tended orchard.

More walkthroughs

Explore it for yourself

Get started on GitHub