Canonical coordination layer for three sibling open-source AI governance projects.
Current coordination baseline: v0.2. Individual project versions remain listed below.
Modern AI systems let interpretation, decision-making, and execution share one trust boundary. That lets probabilistic systems directly own deterministic consequences — the root cause of prompt injection, unsafe automation, context contamination, and authority confusion.
The context-stack separates interpretation from authority: intelligence proposes, governance validates, execution authorizes. Intelligence can suggest anything. Authority stays deterministic.
| Project | Question it answers | Status |
|---|---|---|
| ContextOps | How does an org govern its AI context? | v0.1 active |
| ContextBoundary | Where is data allowed to go? | v0.1 active |
| Sthala | Where does the AI actually run? | v0.1 Alpha |
| Griha | Product layer — inherits all three principles | Proof of concept |
Griha is the product layer above the three governance projects — it inherits ContextOps, ContextBoundary, and Sthala principles in a working system.
All Apache 2.0. All at github.com/kannanokannan.
- GLOSSARY.md — single source of truth for all terminology used across the three projects
- DECISIONS.md — locked cross-project decisions log with dates
- CLAUDE.md — agent briefing for any LLM working across the stack
- STACK_MANAGEMENT.md — operating model for managing the repos as one stack
- REPO_MAP.md — repo ownership boundaries
- stack-maturity-snapshot.md — derivative cross-project maturity view
- RELEASE_CHECKLIST.md — pre-flight and release checks
- okf/ — OKF-style bridge for agent-readable navigation; canonical sources remain the files above
The okf/ folder exposes the stack's control-plane knowledge in an Open Knowledge Format style layout for agents.
This is a bridge, not a new source of truth. The canonical sources remain GLOSSARY.md, DECISIONS.md, REPO_MAP.md, and STACK_MANAGEMENT.md.
Use okf/index.md as the agent-readable entry point for:
- family doctrine
- project roles
- ContextBoundary / Sthala relationship
- Egress Tier terminology
- links back to canonical control-plane files
Shared stack meaning belongs here in context-stack. Repo-local OKF folders, if added later, should contain only repo-owned domain knowledge and should link back to this central bridge.
- No implementation detail (lives in each project repo)
- No roadmap or backlog (lives in Drive)
- No SaaS, no product, no vendor lock-in
ContextOps → governs organisational AI context lifecycle
ContextBoundary → governs where data is allowed to flow (horizontal layer)
Sthala → runtime placement and compliant implementation under ContextBoundary
ContextBoundary is deployment-agnostic. Sthala is a runtime reference that consumes it. Do not conflate the three.