Scintilla Locate
Protocol Engineering Handbook
The canonical reference for engineering Scintilla Locate: governance, Spec‑Driven Development, determinism, economic security, dispute‑grade reproducibility, and an AI‑augmented software factory.
At a glance
How intent becomes safe protocol change
1 Jira intent & acceptance criteria
2 Agents draft proposals (never merge)
3 Deterministic gates build IR + verify
4 Human gates approve spec/plan/risk/release
5 CI proves conformance; release is auditable
Start here
Protocol overview
Understand what Scintilla Locate is, what it optimizes for, and where authority lives.
Open →
Threat model
The adversarial assumptions that drive protocol design and engineering constraints.
Open →
Economic security model
The self‑sizing exposure envelope and why it matters to engineering and governance.
Open →
Engineering system
Governance charter
The constitutional layer: who decides what, and how protocol change is authorized.
Open →
Agent constitution
Hard operational laws for AI agents: determinism boundaries, auditability, least privilege.
Open →
Protocol change process
The required artifacts, gates, and CI evidence to land a protocol change safely.
Open →
AI software factory
Orchestrator architecture
Planner → Builder → Verifier workflows, policy routing, immutable run ledger, and deterministic tooling.
Open →
Jira agent workflows
How Jira drives intent while specs + CI remain the authority and acceptance mechanism.
Open →
Factory architecture diagram
A single diagram that shows the full lifecycle: intent → agents → deterministic gates → humans → release.
Open →