Log inSign up
Straiker
546 posts
Straiker profile banner
@straikerai

Straiker

@straikerai
The Agentic Security Company straiker.ai
straiker.ai
Joined November 2024
81
Following
393
Followers
RepliesRepliesRepostsRepostsMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
  • @straikerai
    Straiker
    @straikerai
    Sep 3
    Let the agent go rogue. Just give it somewhere safe to do it. 🔥 High-risk agents need realistic environments to uncover unexpected behavior. That means giving them access to tools, systems, code, and permissions that let them behave the way they might in the real world. But
    straiker.ai
    Heimdall: Cyber-Hardened Infrastructure for High-Risk Agent Evaluation and Training | Straiker
    Heimdall is a cyber-hardened AI sandbox for high-risk agent training and evaluations, with independent controls for isolation, egress, telemetry, and shutdown.
    2
  • @straikerai
    Straiker
    @straikerai
    Sep 3
    We believe every AI agent needs a kill switch. We also believe every startup needs a pinch-me moment. Ours came two months ago, when we closed our Series A and lit up the Nasdaq Tower. A good reminder in the lights of why we’re here: to prevent agentic chaos and secure the
    Image
    00:00
  • @straikerai
    Straiker
    @straikerai
    Sep 1
    Our research here at @straikerai was just cited by @VentureBeat📣 The article highlights findings from our STAR Labs Threat Report: 91% of successful attacks on productivity agents resulted in silent data exfiltration. Read more ⤵️
    @VentureBeat
    VentureBeat
    @VentureBeat
    Sep 1
    An Azure OpenAI assistant exposed SharePoint data due to a retrieval flaw, highlighting a critical gap in custom pipelines' entitlement checks. 91% of successful attacks on productivity agents ended in silent data exfiltration. venturebeat.com/ai/azure-opena…
    1
  • @straikerai
    Straiker
    @straikerai
    Aug 31
    Welcome Joey Popelka and Antoine Juhasz to Straiker ⚡ 🃏 Joey joins Straiker Labs as an Engagement Manager, bringing a mix of computer science, finance, investing, and a love of poker. 🛡️ Antoine joins as Principal AI Security Sales Engineer for the East, bringing 20 years
    Image
  • @straikerai
    Straiker
    @straikerai
    Aug 28
    We asked Rob van der Veer, Founder of the OWASP AI Exchange, a pretty straightforward question during our latest webinar: Is input filtering actually an effective defense against prompt injection? 🤔 Rob didn’t mince words: “I think it’s one of the least effective approaches.”
    Image
    You Can't Filter Your Way Out of Prompt Injection Attacks | Straiker
    From straiker.ai
Advertisement
Advertisement