# Ecosystem & Roadmap > Current status, version roadmap, and ecosystem fit for ACS and AOS - **Category**: guardrails-acs - **Canonical URL**: https://designpattern.fyi/guardrails-acs/ecosystem-roadmap/ --- ## Description Current status, version roadmap, and ecosystem fit for ACS and AOS ## Additional Notes # Current State, Roadmap, and Ecosystem Fit --- ## Version Status Both the AOS and ACS repositories publish the same staged roadmap: ### v0.1 (Public Preview) — *Current Stage* - **Documentation** — comprehensive guides and reference materials - **Definitions/Schema** — hook specifications, event schemas, verdict objects - **OpenTelemetry/OCSF Observability Definitions** — agent-specific semantic conventions - **AgBOM Requirements** — specifications for dynamic component inventory ### v1 — *Planned* - **Reference Guardian Agent Implementation** — production-ready enforcement engine - **AOS/ACS-to-OpenTelemetry Mapper** — standard integration components - **AOS/ACS-to-OCSF Mapper** — security event conversion tools - **FastMCP Client Instrumentation** — Model Context Protocol integration - **A2A Client Instrumentation** — Agent-to-Agent protocol integration ### v2 — *Planned* - **Full CycloneDX Requirements** — complete AgBOM specification for CycloneDX - **Full SPDX Requirements** — complete AgBOM specification for SPDX - **Full SWID Requirements** — complete AgBOM specification for SWID - **Mapper Implementations** — automated AgBOM generation and management tools ### v3 — *Planned* - **MCP Protocol Extensions** — native deny/modify operations in MCP itself - **A2A Protocol Extensions** — native deny/modify operations in A2A itself - **Protocol-Level Control** — move from external gating to protocol-native enforcement --- ## Active Workstreams Per the ACS site and launch coverage, several workstreams are actively progressing: ### Middleware/Guardian Agent Reference Implementation - **Status:** Active - **Focus:** Production-ready Guardian Agent implementation - **Goal:** Provide a reference implementation that organizations can deploy or adapt ### OpenTelemetry Upstream Contribution - **Status:** Active - **Focus:** Contributing agent-specific semantic conventions to OpenTelemetry - **Goal:** Make agent observability a standard part of the OpenTelemetry ecosystem ### AgBOM Extensions to CycloneDX/SPDX - **Status:** In Progress - **Focus:** Extending existing BOM standards for agent-specific requirements - **Goal:** Enable agent component tracking using industry-standard formats ### MCP/A2A Protocol Integration - **Status:** Planned - **Focus:** Deep integration with Model Context Protocol and Agent-to-Agent protocol - **Goal:** Native protocol-level support for ACS controls ### Identity Workstream - **Status:** Announced - **Focus:** Non-human agent authentication, ephemeral credentials, just-in-time access - **Goal:** Solve the agent identity and privilege management challenge ### Coding-Agent Workstream - **Status:** Announced - **Focus:** Extending ACS hooks to IDE-based agents - **Goal:** Apply ACS controls to code generation, file edits, command execution --- ## Ecosystem Fit ACS explicitly builds on rather than competes with adjacent standards: | Standard | Role Relative to ACS | |---|---| | **MCP** | Tool-calling protocol; ACS adds control hooks around MCP traffic rather than replacing it | | **A2A** | Inter-agent communication protocol; same relationship — ACS adds control hooks around A2A traffic | | **OWASP Top 10 for Agentic Applications** | Risk catalog; ACS is one implementation path for mitigating it (see Section 5) | | **AIVSS (Agentic AI Vulnerability Scoring System)** | Sibling OWASP initiative for scoring vulnerability severity; complements ACS's enforcement focus | | **OpenTelemetry** | Tracing backbone; ACS contributes agent-specific semantic conventions upstream | | **CycloneDX / SPDX / SWID** | Existing BOM standards; ACS extends them for AgBOM rather than inventing a new format | | **EU AI Act / NIST AI RMF** | Regulatory frameworks that specify *what* is required; ACS proposes *how* to implement it technically | ### Integration Philosophy ACS follows a **complement, not replace** philosophy: - **Builds on existing standards** — leverages OpenTelemetry, OCSF, CycloneDX, SPDX, SWID - **Extends rather than invents** — adds agent-specific capabilities to proven frameworks - **Interoperates by design** — works with existing security and observability infrastructure - **Community-driven** — contributes upstream rather than creating parallel ecosystems --- ## Current Adoption Status ### A Snapshot, Not a Verdict As of this writing (July 2026): - **OWASP AOS Repository:** ~53 stars, ~31 forks - **ACS Repository:** ~37 stars, ~9 forks **Important Context:** - Both numbers will be stale by the time you read this - Check the repositories directly for current activity, open issues, and framework support - Early adoption numbers don't reflect production readiness or long-term viability - Public-preview status means active development and rapid evolution ### Framework Support Matrix Check the current framework-support matrix before assuming your stack is covered: - **Native hook support** — some frameworks have first-class AOS integration - **Wrapper-based integration** — some frameworks require manual wrapper integration - **No current support** — some frameworks may not yet have AOS integration **Recommendation:** Verify current support for your specific frameworks before planning implementation. --- ## Community and Governance ### Project Governance Both AOS and ACS are community-governed: - **No single vendor ownership** — Zenity is a lead contributor and sponsor, not the owner - **Open-source licenses** — Apache 2.0 for AOS, MIT for ACS - **Community contributions** — accept contributions from the broader community - **Transparent development** — issues, discussions, and roadmap are public ### Contributing Ways to contribute to the ACS/AOS ecosystem: - **Implementation feedback** — share lessons from production deployments - **Issue reports** — report bugs or gaps in the specification - **Feature requests** - propose new capabilities or improvements - **Documentation** — improve guides and reference materials - **Code contributions** — submit implementations or integrations ### Community Resources Engage with the ACS/AOS community: - **GitHub repositories** — issues, discussions, pull requests - **ACS site** — [agentcontrolstandard.org](https://agentcontrolstandard.org) - **AOS site** — [aos.owasp.org](https://aos.owasp.org) - **OWASP community** — participate in OWASP GenAI Security Project - **Conferences and events** — AI Agent Security Summit and other industry events --- ## Production Readiness Considerations ### Public-Preview Status ACS is currently in public-preview (v0.1.x), which means: - **Actively evolving** — specifications and implementations may change - **Limited production case studies** — few large-scale enterprise deployments documented - **Rapid iteration** — frequent updates based on community feedback - **Breaking changes possible** — later versions may include breaking changes ### Evaluation Before Adoption Before committing to ACS for production use: 1. **Check current framework support** — verify your stack has native or wrapper-based integration 2. **Review issue backlog** — understand known limitations and planned fixes 3. **Evaluate reference implementations** — assess if they meet your requirements 4. **Plan for evolution** — consider how you'll handle specification changes 5. **Engage with the community** — understand the roadmap and contribute feedback ### Complement, Not Replace Consider ACS as part of a broader guardrails strategy: - **Combine with other controls** — use ACS alongside existing security measures - **Defense in depth** — layer ACS with other security controls for robust protection - **Gradual adoption** — start with non-critical agents and expand coverage over time - **Continuous evaluation** — assess effectiveness and adjust your approach as needed --- ## Key Takeaways 1. **Staged roadmap** — clear path from public preview to protocol-level controls 2. **Active development** — multiple workstreams progressing in parallel 3. **Ecosystem integration** — builds on existing standards rather than competing 4. **Early adoption stage** — limited production case studies, active evolution 5. **Community governance** — no single vendor ownership, open-source licenses 6. **Evaluate before committing** — assess framework support and production readiness 7. **Part of broader strategy** — use ACS as part of a defense-in-depth approach --- ## Next Steps - **Limitations & Critique** — understand the current constraints and critical perspectives - **Key Resources** — access official documentation and community links