Loading...
AI agent tooling, tutorials, and the path to making agents first-class software consumers.
REST, MCP, A2A, llms.txt, and OpenAPI — why JarvisSDK supports all five discovery protocols, and how each one serves a different type of AI agent consumer.
Anthropic says build skills, not agents. Here's why skills — portable, testable, composable units of encoded knowledge — are the future of AI tooling, and what that means for how we build.
A hands-on tutorial: create a JarvisSDK module, register it in the catalog, and execute it via the REST API—in under five minutes. Includes TypeScript code samples using @jarvis-sdk/client.
When an AI agent executes arbitrary code from an unknown source, who's responsible for what happens? JarvisSDK's trust scoring, certification pipeline, and sandboxing answer the hardest question in autonomous AI: can you actually trust the tools your agents run?
AI agents today discover tools through hardcoded lists and manual configuration. JarvisSDK is building the universal discovery and execution layer—npm for AI agents—so any agent on any framework can gain capabilities instantly.