End-to-end walkthroughs demonstrating Supergraph engineering pipelines across common scenarios:
- Scientific Debugging: 6-phase root-cause isolation with Serena LSP and
diagnostician. - Wave DAG Feature Implementation: Multi-worker git worktree execution with
supergraph-team. - Full-Stack QA & Anti-Bot Testing: Native Rust CDP (
agent-browser) with 5-dimension ISTQB matrix. - Architectural Refactoring: System contracts, blast-radius mitigation, and independent review.
# 1. Start session with graph context
/supergraph:scan
# 2. Bug cause unknown -> run scientific diagnosis
/supergraph:diagnose
# Output: Falsifiable hypotheses, Serena symbol trace, minimal reproduction
# 3. Create test reproducing failure
/supergraph:tdd
# 🔴 RED: run test -> fails
# 4. Implement minimal fix
# 🟢 GREEN: run test -> passes
# 5. Auto-fix loop + verification
/supergraph:fix
/supergraph:verify# 1. Analyze problem and blast radius
/supergraph:analyze
# 2. Create architecture-aware implementation plan
/supergraph:plan
# Output: docs/supergraph/plans/YYYY-MM-DD-feature.md
# 3. Execute plan via Supergraph Team
/supergraph:execute
# Initializes sg_team_init, spins up isolated worktrees, streams progress to http://127.0.0.1:3333
# 4. Auto-fix tests & lint
/supergraph:fix
# 5. Run E2E & Full-Stack QA (agent-browser)
/supergraph:browser-agent
# Generates docs/qa/test-matrix-<feature>.md and verifies 0 console errors
# 6. Verify and independent review
/supergraph:verify
/supergraph:review# 1. Architectural inspection
/supergraph:architecture
/supergraph:zoom-out
# 2. Scope & plan with embedded contracts
/supergraph:sdd
/supergraph:plan
# 3. Wave execution with isolated git worktrees
/supergraph:execute
# 4. AppSec review before merge
/supergraph:vibesec
/supergraph:review