Cosmos IBC Protocol Adoption: The Neutral Foundation Shift

On October 22, 2026, Anthropic donated the Agent Skills standard to the Agentic AI Foundation (AAIF) under the Linux Foundation. This move places the SKILL.md format alongside the Model Context Protocol (MCP) and AGENTS.md in a neutral governance structure that no single vendor controls. I see this as a strategic pivot from proprietary ownership to open infrastructure, driven by the fact that more than 70 agent platforms had already integrated the standard before the handover. The adoption numbers are undeniable: 1 million skills published and nearly 280 million installed since the standard’s release in December 2025. These figures prove that developers treat skills as essential tools rather than experimental features. The foundation now hosts ten projects, up from three at its inception, including Open Responses from OpenAI and OpenEnv from Hugging Face.

Governance and Technical Scope

The AAIF was co-founded by Anthropic, Block, and OpenAI with support from Google, Microsoft, AWS, Cloudflare, and Bloomberg. Engineers from these companies maintain Agent Skills as individuals, with David Soria Parra and Mahesh Murag serving as Lead Maintainers. A skill is defined as a folder containing a SKILL.md file that teaches an agent a specific task or domain, along with any scripts or reference material it needs. Agents load these files only when a relevant task arises, which allows a single agent to carry hundreds of capabilities without loading them all into context simultaneously. This design minimizes latency and resource usage compared to static system prompts. The proposed roadmap includes versioned releases, a formal proposal process, and security guidance for people who write and install skills. Security remains the most critical gap because a skill can include executable scripts, and with nearly 280 million installs, vetting procedures are still on the roadmap rather than published rules. You should treat third-party skills like code you install: read their scripts before an agent runs them.

Strategic Implications for Developers

Giving away a winning standard is strategy, not generosity. When your format becomes the default that everyone builds on, you win even without owning it. Anthropic’s leverage shifts from controlling the spec to being the company that understands it best and ships the best agent on top of it. Neutral governance removes the biggest objection cautious enterprises raise: the fear of betting on one vendor’s proprietary format. Handing governance to a foundation widens the whole market Anthropic sells into. For developers, building on Agent Skills no longer means betting on Anthropic’s roadmap; it means betting on a shared standard with 70-plus platforms behind it. That is a very different risk calculation. The pattern is clear: MCP handles how agents connect to tools and data, while Agent Skills handles what capabilities an agent can pick up and use. Both sit in the same neutral foundation, creating an interoperable layer that no single company can fence off. Will the maintenance actually remain independent over time? Judge it by the commit history, not the press release.