Skip to main content

The Agent Contract
Roadmap

Evolving the missing layer that bridges protocols to platforms. Vendor-neutral, protocol-agnostic, truly portable.

What Just Shipped

Recently delivered in v0.4.8 — Skills Pipeline & NIST Alignment

Skills Pipeline

  • Reusable capability bundles (prompts, tools, policies) attachable to any agent
  • Skills research, generate, and export workflow
  • NIST CAISI RFI alignment with W3C DIDs, Cedar authorization, and SP 800-53 mapping

MCP & Tooling

  • 10 production MCP tools (validate, scaffold, convert, inspect, generate, publish, list, workspace, diff, migrate)
  • Claude Agent SDK export target
  • Drupal-native OSSA manifest import (ai_agents_ossa module)

Export Targets

  • 23+ export targets including Docker, Kubernetes, LangChain, CrewAI, Cursor, Warp
  • A2A agent-card.json generation
  • Mobile Agent runtime export

Developer Experience

  • 65+ CLI commands
  • Interactive wizard for agent creation
  • 121 test suites

Current Focus: v0.4.x

Specification release, major framework integrations, and developer tooling.

SPECSpecification v0.4

  • Enhanced security definitions for RBAC
  • Standardized tools schema normalization
  • New memory persistence interface
  • Multi-modal input specifications

2/4 complete

Done In Progress Planned

TOOLINGDeveloper Experience

  • Visual VS Code Extension
  • CLI migrate command for LangChain
  • Interactive Agent Builder 2.0
  • CI/CD Validation Actions

1/4 complete

Done In Progress Planned

ECOSYSTEMIntegrations

  • Official Python SDK
  • TypeScript SDK v1.0
  • LlamaIndex Native Support
  • AutoGen Adapter

1/4 complete

Done In Progress Planned

Timeline to v1.0

Honest milestones. No hype. No fake deadlines.

v0.4.x
Current — Alpha
Now

Core specification stable. Security model, multi-agent orchestration, OpenTelemetry tracing. Active community feedback.

2
v0.5.0
Beta
Q3 2026

Memory persistence, full Python SDK, framework migration CLI, CI/CD validation actions. API stability guarantees.

3
v0.6.0
Release Candidate
Q4 2026

Enterprise governance features, policy-as-code, immutable audit trails, cost attribution. Conformance test suite finalized.

4
v1.0.0
Stable
2027

Full backward compatibility guarantee. Stable API surface. Production-grade conformance certification.

What's Next

Scaling to multi-agent orchestration and enterprise governance.

Orchestration Protocol

Building the OSSA contract layer for A2A communication — standardized handoffs and distributed task execution.

  • Standardized handshake protocols
  • Context transfer schemas
  • Reputation and trust scoring

Enterprise Governance

Advanced features for regulatory compliance and organizational control.

  • Policy-as-Code for agents
  • Immutable audit trails
  • Cost attribution standards

Drupal Integration

Bringing AI agent orchestration to the Drupal ecosystem.

OSSA integrates with Drupal through ECA (Event-Condition-Action) workflows, enabling content teams to define and deploy AI agents alongside their CMS without leaving the Drupal ecosystem.

Shipped

ECA-to-OSSA Migration Guide

Convert Drupal ECA workflows to portable OSSA agent manifests.

In Progress

Drupal Module for OSSA

Native Drupal module for managing OSSA agent manifests within the CMS.

Planned

Drupal Recipe Integration

Drupal recipes that scaffold OSSA-powered agents for common CMS workflows.

Research & White Papers

Upcoming deep dives into agent architecture and data science.

IN PROGRESS
Q1 2026

The State of Agent Interoperability

A quantitative analysis of 50+ agent frameworks and the fragmentation cost for enterprises.

PLANNED
Mar 2026

Defining Agent "Cognitive Load"

Proposed metrics for measuring agent complexity and reasoning overhead in production systems.

PLANNED
Q2 2026

Security Patterns for Autonomous Agents

A comprehensive guide to threat modeling and defense strategies for multi-agent systems.

Submit a Topic

Have a research idea or data science topic related to agents?

Open an Issue →

Help Shape the Standard

OSSA is community-driven. Your input directly influences our roadmap and priorities.