Skip navigation
AI agents are getting powerful enough to need their own security layer
0:00 / 5:29

AI agents are getting powerful enough to need their own security layer

Alba

@albacnt43K followers

Visit on X
136 likesDownloadOriginal post

19,824 views

AI agents are getting powerful enough to need their own security layer NVIDIA just open-sourced OpenShell It lets you sandbox agents, control what they can access and monitor what they’re doing Watching an agent work inside these boundaries is pretty wild

19 Comments

Comments from the original post on X (Twitter)

  • yuulia ✟'s profile picture

    @yuliaisc

    we're going to need way more of this as agents get more autonomous

    1 like

  • liam.'s profile picture

    @liambraus

    the more capable agents get, the more important this becomes

    1 like

  • marcus's profile picture

    @marcusyul

    Este vídeo pinta bien.. gracias amigo

    1 like

  • Buswe's profile picture

    @buswe_com

    The network side is where it shows most: with an egress allowlist, a confused agent simply cannot reach anywhere it should not.

    1 like

  • Nadia's profile picture

    @nadia7ro

    the idea of being able to quarantine an agent is wild

    1 like

  • naldz.'s profile picture

    @naldzyul

    giving agents access to computers without proper isolation sounds terrifying

    1 like

  • Javeriya Ahsan's profile picture

    @AhsanJaveriya

    Agents need permissions and boundaries, not just better reasoning.

    1 like

  • robin.'s profile picture

    @robinstetic

    this is probably one of the least talked about problems with agents

    1 like

  • Pelis Free (Follow)'s profile picture

    @Pelis777Gratis

    autonomous agents + no restrictions is a scary combination

    1 like

  • tobarra's profile picture

    @txbrraa

    imagine an agent running for hours without anyone watching it

    1 like

  • ryu.'s profile picture

    @romanftp

    AI agents needing security layers is honestly a pretty big shift

    1 like

  • FeNiX_MoDz's profile picture

    @FeNiX_MoDz

    agent security is about to become its own entire industry

    1 like

  • Reyad Yosef's profile picture

    @YosefReyadjo

    where do the API keys live in this setup? if they're inside the sandbox, a prompt-injected agent can still just read them

    1 like

  • Ryoru 𒌐's profile picture

    @Ryorugg

    sandboxing agents is going to become standard

    1 like

  • Peliculas Gratis (Follow)'s profile picture

    @lherrburhn

    this is going to matter a lot more sooner than people think

    1 like

  • toba's profile picture

    @999toba

    sandboxing might be one of the most important pieces of agent infrastructure

    1 like

  • Aleksandar Janca's profile picture

    @datachad

    I still keep anything irreversible out of the agent completely, sending and paying go through plain if/then

    1 like

  • Anderson's profile picture

    @ItsGoharr

    agents need a security layer, ship them without one first

    1 like

  • SunsetSyntax's profile picture

    @sunsetsyntax

    Sandbox demos look fine until the policy is prompt-only. Tip: put the deny list in the runtime that wraps the process, and make a restart restore the same policy from disk - otherwise a crash is an escape hatch.

    1 like