Skip to content

Latest commit

 

History

History
219 lines (165 loc) · 10.4 KB

File metadata and controls

219 lines (165 loc) · 10.4 KB
title Roadmap
description SINT Protocol development roadmap — from npm publish to enterprise adoption.
sidebarTitle Roadmap

This roadmap is based on what we've built, what the market needs right now, and the fastest path to adoption and revenue. Updated April 2026.


✅ Phase 0: Foundation (Complete)

Everything built and verified:

Deliverable Status Details
SINT Protocol core ✅ Complete 24 packages, 1,363 tests, 67 build tasks
Policy Gateway ✅ Complete 32 REST endpoints, intercept pipeline
Capability Tokens ✅ Complete Ed25519, attenuation-only delegation, max depth 3
Evidence Ledger ✅ Complete SHA-256 hash chain, proof receipts, SIEM export
12 Protocol Bridges ✅ Complete ROS2, MCP, gRPC, MAVLink, A2A, MQTT, OPC-UA, IoT, OpenRMF, Economy, Swarm
3 SDKs ✅ Complete TypeScript, Python (1,871 lines), Go (220 lines)
sintctl CLI ✅ Complete Token, ledger, approvals, policy management
SINT Console ✅ Complete 31 features, 60+ feature flags, WebSocket gateway
Visual Workflow Builder ✅ Complete 13 node types, topological execution, n8n integration
Goal Hijack Detector ✅ Complete 25+ patterns, confidence scoring
Security Audit ✅ Complete 0 vulnerabilities, all 8 attack surfaces verified
Documentation ✅ Complete 19 pages on docs.sint.gg
SINT Avatar ✅ Complete 3D avatar, ElevenLabs lipsync, 12 expressions
CMO Operator ✅ Complete 18 skills, video→content pipeline
SINT Outreach ✅ Complete LinkedIn automation, BrightBeam pilot active

🚀 Phase 1: Ship & Open (April 2026)

Goal: Make SINT Protocol installable by anyone. Get the first external integrations.

Publish to npm registry: `@sint/core`, `@sint/gate-capability-tokens`, `@sint/gate-policy-gateway`, `@sint/gate-evidence-ledger`, `@sint/persistence`, `@sint/bridge-mcp`, `@sint/client`, `@sint/bridge-ros2`
**Blocked:** Needs NPM_TOKEN or interactive `npm login`.

**Impact:** Unlocks all downstream integration work.
Write `@sint/langchain` — a LangChain callback handler that wraps every tool call in a SINT capability token check.
```python
from sint import SintGovernanceHandler
chain = agent.with_config(callbacks=[SintGovernanceHandler(gateway_url)])
```

**Target:** LangChain community packages, blog post on LangChain blog.
Write `@sint/crewai` — CrewAI task decorator that enforces capability tokens before execution.
**Target:** CrewAI plugin registry.
"Show HN: Open-source safety governance for AI agents — capability tokens, policy gateway, evidence ledger"
Position: "We built what Microsoft's Agent Governance Toolkit does, plus physical AI safety (robots, drones, industrial)."
Submit SINT Protocol as reference implementation to NIST CAISI's AI Agent Standards Initiative comment period.
Position: SINT covers all 10 OWASP Agentic Top 10 risk categories with working code.
Contact OWASP Agentic Security Initiative team. Get SINT listed as reference implementation for their compliance framework.

Timeline: 2-3 weeks after npm publish unblock.


🔧 Phase 2: Framework Middleware (May 2026)

Goal: Become the default governance layer for popular agent frameworks.

Integration Framework What We Build
@sint/langchain LangChain / LangGraph Callback handler + chain interceptor
@sint/crewai CrewAI Task decorator + crew-level policy
@sint/autogen Microsoft AutoGen Agent runtime wrapper
@sint/google-adk Google ADK Plugin for A2A governance
@sint/openai-agents OpenAI Agents SDK Tool middleware
@sint/openclaw OpenClaw Policy provider plugin

Each integration is a thin wrapper (~200-500 lines) that:

  1. Intercepts tool calls / agent actions
  2. Validates against SINT capability tokens
  3. Logs to evidence ledger
  4. Returns approve/deny/escalate

Success metric: 3+ framework integrations published, 100+ npm downloads/week.


🤖 Phase 3: Physical AI Partnerships (June-July 2026)

Goal: First real-world robotics integration. Prove the thesis that AI agents controlling physical systems need governed middleware.

Target What They Do SINT Value Approach
RobotecAI (RAI) ROS2 agentic robotics ROS2 bridge + capability tokens PR to rai repo with SINT safety layer
Dronecode / PX4 Open-source drone platform MAVLink bridge + geofence enforcement Plugin for QGroundControl
GrayMatter Robotics LA-based factory AI Force/velocity constraints + audit trail Direct sales (local neighbor)
NVIDIA IsaacSim Robot simulation T1 Sandbox → T4 Autonomous tier progression IsaacSim extension
Open-RMF Multi-robot fleet mgmt OpenRMF bridge + zone governance Community contribution

First target: RobotecAI — they have a vendor-agnostic ROS2 framework with zero safety governance. Our @sint/bridge-ros2 drops in directly.

Success metric: 1 working robotics integration with external project.


🏢 Phase 4: Enterprise & Compliance (Q3 2026)

Goal: Position SINT as the compliance answer for enterprise AI agent deployments.

OWASP Agentic Top 10 Compliance Mapping

OWASP Risk SINT Mitigation Package
AG01: Excessive Agency Capability token scoping @sint/gate-capability-tokens
AG02: Inadequate Sandboxing T1 Sandbox tier + capsule sandbox @sint/engine-capsule-sandbox
AG03: Unrestricted Resource Access Constraint enforcement (force, velocity, geofence) @sint/gate-policy-gateway
AG04: Insufficient Credential Handling Ed25519 identity, no credential sharing @sint/gate-capability-tokens
AG05: Prompt Injection Goal hijack detector (25+ patterns) @sint/gate-policy-gateway
AG06: Inadequate Audit SHA-256 evidence ledger + proof receipts @sint/gate-evidence-ledger
AG07: Lack of Human Oversight T2 approval gates + SSE/WebSocket approval queue Gateway server
AG08: Insufficient Error Handling Circuit breaker (e-stop invariant I-G2) @sint/gate-policy-gateway
AG09: Supply Chain Risk Ed25519 supply chain verifier @sint/gate-policy-gateway
AG10: Insufficient Monitoring CSML anomaly scoring + risk stream @sint/gate-evidence-ledger

Enterprise Features to Build

Feature Description Priority
Redis-backed rate limiting Multi-instance rate limit state 🟡 Medium
PostgreSQL evidence store Production-grade persistence 🟢 Ready (adapter exists)
SSO integration SAML 2.0 / OIDC for enterprise auth 🟡 Medium
SIEM connectors Splunk, Datadog, Elastic integration 🟡 Medium
Compliance reports Auto-generated OWASP / SOC2 evidence reports 🔴 High
Multi-tenant isolation Namespace-level resource separation 🟡 Medium

Success metric: 1 enterprise pilot ($5K-$15K/mo), OWASP compliance report generator shipped.


🌐 Phase 5: Protocol Network (Q4 2026)

Goal: Decentralized agent governance — multiple organizations running SINT gateways and trusting each other's capability tokens via the Open Agent Trust Registry.

Component Status Next Step
Trust Registry ✅ 8 issuers registered Open registration, governance voting
Token delegation ✅ Cross-agent delegation working Cross-organization delegation
A2A bridge ✅ Agent-to-agent protocol Multi-gateway federation
Economy bridge ✅ Budget tracking Cross-org settlement

Vision: Agent A (governed by Org 1's gateway) delegates a capability token to Agent B (governed by Org 2's gateway). Both gateways verify the token chain. Evidence is logged to both ledgers. The Trust Registry validates both issuers.


Revenue Timeline

Quarter Source Target
Q2 2026 BrightBeam pilot closes ($5K) $5K
Q2 2026 GrayMatter Robotics intro (LA) Pipeline
Q3 2026 Enterprise compliance pilot (1 client) $10-15K/mo
Q3 2026 Framework integration consulting $5-10K/mo
Q4 2026 Protocol SaaS (hosted gateway) $2-5K/mo/customer

What Makes SINT Different

SINT Protocol Microsoft AGT Proofpoint AI Security
Physical AI ✅ ROS2, MAVLink, OPC-UA, MQTT ❌ Software only ❌ Software only
Open source ✅ MIT ✅ MIT ❌ Proprietary
Capability tokens ✅ Ed25519, attenuation-only ❌ Policy-only ❌ N/A
Evidence ledger ✅ SHA-256 hash chain ❌ Logging only ❌ Proprietary
OWASP coverage ✅ 10/10 ✅ 10/10 Partial
Framework support 🔜 Coming (Phase 2) ✅ LangChain, CrewAI, ADK ❌ Standalone
Console UI ✅ 31 modules ❌ CLI only ✅ SaaS dashboard
Price Free (MIT) + hosted SaaS Free (MIT) $$$$ enterprise
Microsoft AGT shipped April 2, 2026. They have first-mover in enterprise frameworks (LangChain, CrewAI, ADK integrations). Our differentiation is physical AI + capability tokens + evidence ledger + console UI. We must ship framework integrations fast (Phase 2) before they absorb that market.

Contributing

Want to help? The highest-impact areas right now:

  1. Framework integrations — Write a SINT wrapper for your favorite agent framework
  2. Bridge adapters — Add new protocol bridges (MQTT 5.0, WebSocket, Bluetooth LE)
  3. Console features — Enable and build out flagged features (60+ flags, many stubbed)
  4. Trust Registry — Register as an issuer, add your organization
  5. Documentation — Tutorials, guides, video walkthroughs
View source and contribute Join the community