K8s is great at keeping things alive. It's not built to reason about whether the thing that's alive is actually working correctly. Agent infra needs to handle rollback at the logic level, not just the container level.
We're thinking a lot about how we could provide a "Convex" like experience where we guide your coding agents to set up your agents in a way that maximizes the ability to rollback. For example, instead of continuously taking action, it's better that agents gather all required context, do the work needed to make a decision (research, synthesize, etc.), and then only take action in the real world at the end. If an agent did bad work, then this makes it easy to rollback to the point where the agent gathered all the context, correct it's instructions, and try again
One thing we've run into: the filesystem abstraction works well for code artifacts, but when agents interact with external services (APIs, databases), you need a separate "action journal" that logs intended mutations before executing them. This gives you rollback even for side effects that aren't file-based.
The harder unsolved problem is multi-agent state coordination. When Agent A writes to a shared workspace and Agent B reads it, the filesystem gives you structural persistence but not semantic ordering. Have you thought about adding lightweight event ordering (like a simple monotonic sequence) to the filesystem layer? Without it, agents can read stale state even when files are fresh.
Congrats on the launch — the decoupled storage model is the right primitive.
> Our biggest pain point with hosting agents was that you'd need to stitch together multiple pieces: packaging your agent, running it in a sandbox, streaming messages back to users, persisting state across turns, and managing getting files to and from the agent workspace.
The k8s ecosystem already handles most this and your agent framework the agent specifics. What you are talking about is valid, though a different axis imo. Quality and guardrails are important, but not discussed by OP.
If you repurpose k8s with ephemeral volumes or emptyDir, a sidecar, you'll likely get predictable ops and avoid vendor lock-in. Expect more operator work, fragile debugging across PVCs and sidecars, and the need to invest in local emulation or a Firecracker or gVisor sandbox if you want anything like laptop parity.
The hype is so large with the CLI coding tools I got FOMO, but as you were saying in that thread, I see no tangible improvement to the value I get out of AI coding tools by using the CLI alone. I use the CLI in VS Code, and I use the chat panel, and the only thing that seems to actually make a difference is the "context engineering" stuff of custom instructions, agent skills, prompt files, hooks, custom agents, all that stuff, which works no matter which interface you use to kick off your AI coding instructions.
Would be curious to hear your thoughts on the topic all these months later.
The reasons are (1) it's faster to do admin work like naming or deleting old sessions (2) I have not gotten the remote setup to work yet (haven't tried) but I do want to use it somewhere
But yeah, it's gotten worse, the latest I recall is a new diff viewer for AI in the terminal (I already have git and lazygit)
One thing I took from ATProto is a strong belief that user agency and choice are the penultimate design criteria. To those ends, I think that any agentic tooling needs to support the majority of users' choice about how to interact with it (SDK, API, CLI, TUI, IDE, and Web). My custom agent is headed that way anyhow, because there are times where I do want to reach for one of them, and it's easier to make it so with agents working on their own codebase (minus vscode because the testing/feedback I haven't figured out yet)
Agents run on infra, they have network connectivity, they have ACLs and permissions that let them read+write+execute on resources, they can interact with other agents.
To manage them from both an infra and security perspective, we can use the existing underlying primitives, but it's also useful to build abstractions around them for management, kind of like how microservices encapsulate compute+storage+network together.
I think of agents as basically microservices that can act in non-deterministic ways, and the potential "blast radius" of their actions is very wide. So you need to be able to map what an agent can do, and it's much easier to do that if there are abstractions or automatic groupings instead of doing this all ourselves.
The monitoring problem alone is closer to fraud detection than traditional APM. You're not looking for "is this thing up," you're looking for "is this thing subtly wrong in a way that compounds over the next 10 steps."
Most sandboxing approaches — including this one — optimize primarily for isolation from the host: prevent the agent from escaping, limit what it can touch. That solves the runaway agent problem.
But there's a second axis: observable execution for human collaborators. When an agent modifies a codebase or runs a research task, a teammate often needs to watch it happen in real time, intervene before it commits a wrong turn, or audit what actually ran. Async logs and artifact outputs don't cover this well.
We've been building Cyqle - https://cyqle.in -- (disclosure: I work on it) from that angle — cloud desktop sessions where agent runs are shared live with whoever needs visibility. Isolation is at the VM/session level rather than syscall granularity. Different tradeoff: you give up process-level permission mapping, you gain real-time collaborative access to the running environment — watch, intervene, hand off.
The use cases probably don't overlap much with Terminal Use (async batch filesystem agents clearly want deep process isolation + lifecycle APIs). But this thread made me think "agent environments" is actually several distinct problem spaces: async autonomous execution, interactive human-supervised sessions, team-observable debugging runs. The right primitives look very different depending on which you're solving.
tl;dr, I don't think the shovel analogy holds up for most of the Ai submissions and products we see here.
I think Kata containers with Kubernetes is an even better sandboxing option for these agents to run remotely.
Shameless plugin here but we at Adaptive [1] do something similar.
The permissions issues you mention are handled by SA/WIF and the ADK framework.
Same question to OP, why do you think I need a special tool for this?
We don't need to rebuild everything just for agents, except that people think they can make money by doing so. YC has disappointed me of late with the lack of diversity in their companies. I suspect the change in leadership is central to this.
eg. I already run Kubernetes