{"slug": "closing-the-execution-gap-a-series", "title": "Closing the execution gap: a series", "summary": "Developer Arun Raghunath launched jhansi.io, an open-core cloud sandbox designed to safely execute AI-generated code in production. The platform addresses the gap between AI writing code and running it securely, with a five-part series covering persistent sandboxes, dependency management, isolation, secrets handling, and audit logging.", "body_md": "Every AI coding tool can write Python — Cursor, Claude Code, Windsurf. None of them can run it safely in production.\n\nThat gap between \"AI wrote the code\" and \"the code ran safely\" is exactly what I'm building [jhansi.io](https://jhansi.io) to close.\n\nThis series documents the journey. One layer of the problem at a time.\n\nWhen AI generates code, four things still stand between you and prod:\n\n`eval()`\n\nyour way through an audit.jhansi.io is the missing `run()`\n\nfor AI-generated code. Open core, cloud sandbox, built to close each part of the gap — layer by layer.\n\n**Part 1 — Persistent sandboxes**\n\nWhy \"ephemeral\" breaks debugging, state, and compliance. The case for giving every AI a home directory.\n\n→ [Read Part 1](https://dev.to/thearun85/the-case-for-persistent-sandboxes-in-ai-code-execution-3158)\n\n**Part 2 — Dependency management** *(coming soon)*\n\nDetecting, installing, and locking deps across Python, Node, Go, and Java. With SBOMs and policy built in.\n\n**Part 3 — Isolation** *(coming soon)*\n\nWhat \"hard isolation\" actually means. Containers, Firecracker, zero trust networking, and the metadata service attacks you haven't thought of yet.\n\n**Part 4 — Secrets** *(coming soon)*\n\nKernel-level proxies. AI can call Stripe without the key ever entering the sandbox.\n\n**Part 5 — Audit** *(coming soon)*\n\nWho ran what, when, with which prompt. Hash-chained logs that satisfy auditors, not just engineers.\n\nBuilding this in public. Follow the series on [Dev.to](https://dev.to/thearun85/closing-the-execution-gap-a-series-3490), [Linkedin](https://www.linkedin.com/posts/arun-raghunath_run-ai-generated-code-safely-activity-7469098788822093824-oKzG?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAOmdQMBVGWSljvWa9sZSYfndPCZGwXbz0M), and [X](https://x.com/thearun85/status/2063334004615528556?s=20).\n\nCode is Apache 2.0 at [github.com/jhansi-io](https://github.com/jhansi-io).", "url": "https://wpnews.pro/news/closing-the-execution-gap-a-series", "canonical_source": "https://dev.to/thearun85/closing-the-execution-gap-a-series-3490", "published_at": "2026-06-06 18:51:07+00:00", "updated_at": "2026-06-06 19:11:52.612562+00:00", "lang": "en", "topics": ["ai-infrastructure", "ai-safety", "ai-tools", "ai-products", "ai-startups"], "entities": ["Cursor", "Claude Code", "Windsurf", "jhansi.io", "Firecracker", "Stripe", "Dev.to"], "alternates": {"html": "https://wpnews.pro/news/closing-the-execution-gap-a-series", "markdown": "https://wpnews.pro/news/closing-the-execution-gap-a-series.md", "text": "https://wpnews.pro/news/closing-the-execution-gap-a-series.txt", "jsonld": "https://wpnews.pro/news/closing-the-execution-gap-a-series.jsonld"}}