---
title: "Discussion on \"The current state of security for autonomous agents in the cloud\" | Hashnode"
description: "Discussion on \"The current state of security for autonomous agents in the cloud\". Security controls and best practices for architecting the runtime: isolation, identity, egress, credentials, and a tool broker outside the model."
url: "https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [The current state of security for autonomous agents in the cloud](https://nebulablogs.com/the-current-state-of-security-for-autonomous-agents-in-the-cloud?utm_source=hashnode&utm_medium=feed)

Security controls and best practices for architecting the runtime: isolation, identity, egress, credentials, and a tool broker outside the model.

Sep 20 · 13 min read · nebulablogs.com  
Read the article

-   [#cybersecurity](https://hashnode.com/tag/cybersecurity)
-   [#cloud-computing](https://hashnode.com/tag/cloud-computing)
-   [#ai](https://hashnode.com/tag/ai)
-   [#security](https://hashnode.com/tag/security)
-   [#devops](https://hashnode.com/tag/devops)

## Written by

[Nivethitha Kumaravelu](https://hashnode.com/@eternalnebula)

@eternalnebula · Poland

  
[View profile](https://hashnode.com/@eternalnebula)

Senior Cybersecurity Professional

## Responses 10

Markdown supported · ⌘↵ to post

[Julian Neagu](https://hashnode.com/@julianneagulab)

[Sep 23](https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80/comment/6ab3a8cce70e7b044bcee0f1)

The session-to-user binding bit stood out to me. It’s easy to assume the sandbox solves everything, but identity can still be the weak link.

[BotSailor](https://hashnode.com/@botsailor)

[Sep 21](https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80/comment/6ab10bd7d37697875e76bc4f)

This is an important conversation as AI agents move from experiments into real-world cloud environments. The biggest shift is that we are no longer only securing applications, we are now securing systems that can make decisions, access resources, and take actions on behalf of users.

What I appreciate about this discussion is the focus on fundamentals: identity, isolation, permissions, monitoring, and controlled access. These principles may sound familiar from traditional security, but autonomous agents make them much more critical because a small mistake can scale much faster.

The future of AI agents will not only depend on how intelligent they become, but also on how responsibly we design the boundaries around them. Trust will come from transparency, strong governance, and giving agents only the access they truly need.

Great article. Building powerful AI is exciting, but building secure and reliable AI is what will allow this technology to become part of everyday systems with confidence.

[Nivethitha Kumaravelu](https://hashnode.com/@eternalnebula)

[Sep 21](https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80/comment/6ab10bd7d37697875e76bc4f#reply-6ab13faac5594318117bc07d)

Indeed. It reminds me of the Spiderman quote "With great power comes great responsibility". The security and reliability part is critical for the adoption of these AI systems.

[Mateo Ruiz](https://hashnode.com/@mateoruiz)

[Sep 21](https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80/comment/6ab0dbc6893ba9f03d59d4aa)

The “model proposes, policy disposes” framing is probably the most important architectural distinction here. Once tool calls cross into credentials, network access, or destructive side effects, the model should be treated as an untrusted decision source not the authorization layer.

One additional production concern is making the action broker the single observable choke point. At IT Path Solutions, I’d want every tool invocation to carry the agent identity, user/session context, policy decision, credential scope, and resource being touched. That turns an agent failure from “the model did something strange” into an auditable security event.

The kill-switch sequence is equally important. A control that exists only in documentation isn't really a control until the team has exercised identity revocation, session destruction, credential revocation, and spend freezing as one operational path.

[Nivethitha Kumaravelu](https://hashnode.com/@eternalnebula)

[Sep 21](https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80/comment/6ab0dbc6893ba9f03d59d4aa#reply-6ab13e9216bf3c09031cf584)

The solution you implemented seems pretty solid!

[Ahmet Özel](https://hashnode.com/@ahmetozel)

[Sep 21](https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80/comment/6ab0cbed5c6d4b6d7ac946d6)

Treating prompt injection as intrinsic and then asking what a hijacked plan has to get through is the right way round, and it is the framing most agent-security writing gets backwards by starting at the prompt.

Putting the tool broker outside the model is the load-bearing piece, because it is the only control that still holds when the model is fully compromised.

One addition to the egress plane: deny-by-default is necessary, but the allowlist erodes, and every allowed domain is an exfiltration channel for anything the agent can read. Worth reviewing that list on a schedule rather than only at the moment something gets added to it.

[Nivethitha Kumaravelu](https://hashnode.com/@eternalnebula)

[Sep 21](https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80/comment/6ab0cbed5c6d4b6d7ac946d6#reply-6ab13e3598927792731e39c4)

The egress is critical indeed. Even benign egress can be used as a proxy for exfiltration, as you mentioned. Reviewing the list and monitoring the tool calls/audit logs should help.

[Ahmet Özel](https://hashnode.com/@ahmetozel)

[Oct 4](https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80/comment/6ab0cbed5c6d4b6d7ac946d6#reply-6ac2963b8e7de5c668c30619)

Monitoring tool calls helps, but I would also capture whether an allowed request changed its destination after a redirect. Reviewing only the initial hostname can leave the effective egress path invisible, especially when an otherwise benign endpoint forwards elsewhere.

A focused audit test could send an allowed request through a redirect and require the final destination to receive its own policy decision. Recording the method and destination at each hop would make the review you suggest more concrete, without relying on the model's description of what it intended to send.

[Mihai\_LeanZero](https://hashnode.com/@mihaileanzero)

[Sep 20](https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80/comment/6ab0620eb49ae0e425c25458)

The session-to-user binding point is the one I'd push on hardest, because I've watched a milder version of the same failure mode outside any cloud agent runtime: a shared credentials file sourced by whichever script needs a key that day. Each script only touches the one key it actually uses, but because they all live in the same file, any of them technically could read the rest. Nothing enforces the narrower scope, the discipline just happens to hold so far. Scaled up to a session that spawns a sub-agent mid-task, that's exactly the god-key problem you're describing: does the platform re-issue a narrower credential per hop, or does the whole chain just inherit whatever the first session was granted?

[Nivethitha Kumaravelu](https://hashnode.com/@eternalnebula)

[Sep 20](https://hashnode.com/posts/the-current-state-of-security-for-autonomous-agents-in-the-cloud/6ab0602e91fdd25260202a80/comment/6ab0620eb49ae0e425c25458#reply-6ab06638d760839cd9fe0072)

Good point!! The platforms should ideally re-issue a narrower credential to comply with the security best practices (the principle of least privilege).
