Real Results

Tested in production, not in demos. Everything stated is backed by real code.

Real case: 1 developer, Flutter + Supabase stack. Without SpecBox the estimate was 2-3 months. With the full pipeline: 12 days.

22
User Stories
139
Use Cases
12
Days

The Engine in Numbers

108
MCP Tools
Atomic operations to automate the entire pipeline
13
Agent Skills
Complete auto-discoverable workflows
12
Agents
Each with a defined role in the pipeline
11
Hooks
Automatic enforcement, not warnings
171+
Tests
Engine's own coverage
5
Infra Services
Supabase, Neon, Stripe, Firebase, n8n
3
PM Backends
Trello, Plane, or local without API

Quality Metrics

✅ Coverage ≥ 85%
🔄 Self-healing active
📦 20+ projects managed
🛡️ Zero critical bugs

Why HARD BLOCKS Exist

"9 Use Cases implemented without the pipeline. No traceability. No evidence. The Trello board empty. That was the day the anti-main guard was born."

Now, spec-guard.sh blocks any write to src/ without an active UC. commit-spec-guard.sh prevents commits to main. These are BLOCKING hooks — not warnings you can ignore.

The spec is the contract. The engine enforces it.

The Engine in Action

/prd output

$ /prd "Staff management system"

PRD generated: 4 US, 12 UC, 38 AC

Quality Gate: PASS (38/38 testable)

Trello Board: 12 cards created

✓ doc/prd/PRD_staff.md (2.4 KB)

/implement output

$ /implement US-01

[AG-08] Quality Gate: GO ✓

[AG-09a] 38/38 AC tested

[AG-09b] Verdict: ACCEPTED

→ PR #47 merged, UC-012 complete