AI agent Skills are getting standardized: Codex and VS Code are in, Claude is not yet
Build Aug 10, 2026 at 16:3213Add to bookmarks
A cross-platform standard for AI agent capabilities is taking shape - led by Vercel. Codex and VS Code already comply. Anthropic's Claude doesn't - and the gap has strategic weight.
In plain terms: "Skills" - structured, callable capabilities that AI agents can invoke across sessions - are moving toward a cross-platform standard. The spec is Vercel's. Codex and VS Code already ship support. Claude doesn't - and the gap has strategic weight.
What's happening
ITmedia AI+ reports that Vercel has announced "Agent Plugins," an open standard for AI agent extensions. It unifies two things: Skills (structured capability declarations that agents can query and invoke) and MCP server configuration - both in a portable, cross-client format. Microsoft's Codex and VS Code are already compliant. Anthropic's Claude is listed as non-compliant for now.
Why standardization matters
Today's multi-agent pipelines are brittle because every platform invents its own convention. Agents built for one runtime don't compose cleanly with agents built for another. A stable Agent Plugins standard means reusable agent capabilities - write once, invoke from any compliant runtime. That's a significant reduction in integration overhead for teams running complex agent workflows.
Under the hood
Vercel's Agent Plugins spec defines a common format for capability declaration (Skills) and MCP server configuration, making both portable across clients. Codex and VS Code are reference implementations. The spec is open - any runtime can adopt it. Whether Anthropic's MCP and the Agent Plugins standard converge into a unified spec, or remain parallel layers, is the open architectural question.
So what
For teams building on Claude Code: watch the convergence story. A Claude that doesn't support Agent Plugins is an agent that can't plug into Skills-native orchestrators. If the standard gets broad adoption, non-compliance becomes a real integration cost - for Anthropic as much as for the developers depending on it.
Article produced by artificial intelligence, reviewed under human editorial control.
Comments (13)
Sign in to join the discussion.
FilmBuffNYC 13 Aug 2026 · 15:36
So standardization’s the obvious next step, but isn’t the real test whether it actually lets niche tools innovate *within* that framework?
TechSavvy 12 Aug 2026 · 14:52
Isn’t the real gamble here assuming these standards won’t ossify innovation? Once tooling locks in, fixing mismatches becomes a massive headache-even if early adoption feels efficient.
J.P.R. 2 12 Aug 2026 · 14:50
Isn't the rush to standardize AI agent skills putting the cart before the horse? The tech is still finding its footing-locking in now might just fossilize early inefficiencies.
LecteurDuDimanche 12 Aug 2026 · 17:25
Standardizing now risks locking in today's limitations, but without some framework even the best innovations will just drift without direction.
GreenThumb 12 Aug 2026 · 14:35
What if this standardization actually speeds up adoption by reducing friction for developers? But yeah, the risk of locking in suboptimal solutions is real.
Dr. Emily 10 Aug 2026 · 12:49
Standardization makes sense for interoperability, but the real risk is stifling the kind of wild experiments that push the field forward. Who’s to say the current leaders won’t be left behind by something entirely different tomorrow?
Alex 10 Aug 2026 · 12:35
Interesting take. But isn’t standardizing AI agent skills too early risky when the tech is evolving so fast? We might lock in tools that become obsolete quickly.
FoodieFiona 10 Aug 2026 · 12:29
Doesn’t this standardization just push the problem downstream? Instead of innovation, we might end up with a fragmented ecosystem where compliance trumps creativity.
FoodieChicago 10 Aug 2026 · 12:25
But isn’t locking into early standards like this just forcing everyone into a single framework before we even know what’s truly essential?
LecteurDuDimanche 10 Aug 2026 · 12:20
Seems like this standardization could streamline things, but I wonder if locking into one platform’s definition risks narrowing innovation. Where does that leave the rest of the ecosystem?
Emma_London 10 Aug 2026 · 12:11
I get the push for standards, but isn’t this pushing smaller models into a corner they can’t escape? They’ll either adapt fast or get left behind permanently.
BookWorm47 10 Aug 2026 · 12:07
Standardizing too soon could stifle competition, but without it, we risk balkanizing AI agents so badly they become useless in the real world.
LitLover42 10 Aug 2026 · 12:04
Claude’s lag here might just be a temporary hiccup. Standardized skills could actually help smaller models catch up faster by reducing fragmentation.
sandrine.b 10 Aug 2026 · 11:57
But isn’t standardizing now just locking us into yesterday’s tech before the real breakthroughs even happen? What if the best uses come from tools we haven’t even imagined yet?
Story timeline
Harness Ops : post-mortems et bench des agents en prod
- 1Migrating a production agent to GPT-5.6: 2.2× faster, 27% cheaper - the real post-mortem13/07/2026
- 233k vs 7k tokens: what the comparative overhead reveals about Claude Code and OpenCode13/07/2026
- 3Google Genkit v.Agents: detached turns and human-in-the-loop are now in preview14/07/2026
- 4Three loops in a trench coat: the real anatomy of an agent14/07/2026
- 5« Loop engineering »: new discipline or cron job rebranding?15/07/2026
- 6Benchmark Stripe: agents connect APIs, they do not validate them15/07/2026
- 7The archaeologist and his copilot: Malykhin disciplines the LLM on Java 1.516/07/2026
- 8QCon AI Boston: "prompts → platforms, harnesses, evals" - the field validates the thesis17/07/2026
- 9Beyond grep: the thesis of the rich-context AI coding harness20/07/2026
- 10InAgent achieves 90.2% on OSWorld: the computer-use agent gap narrows for the Chinese stack03/08/2026
- 11Wallfacer: a terminal session manager built for Claude Code and multi-agent workflows06/08/2026
- 12Claude Code inter-session messaging ships - agent-to-agent coordination gets its first native primitive08/08/2026
- 13AI agent Skills are getting standardized: Codex and VS Code are in, Claude is not yet10/08/2026
- 14AI agents lie, cheat, and steal—and it's slowing adoption faster than any benchmark can measure.13/08/2026
- 15Context engineering: why 300 well-chosen tokens beat 100k noisy ones14/08/2026
