Detection for all of your clouds - identity providers, Iaas, Saas, Paas and more.
- Tweets 411
- Following 367
- Followers 1,053
- Likes 118
341+ malicious skills. Credential stealers disguised as a weather app. Command-and-control infrastructure built between AI agents. P0 Labs found this months before OpenClaw security became a mainstream conversation.
hubs.la/Q04q5r_Q0
1
1
401
P0 Labs found Moltbook's AI agent credentials sitting exposed. "You could grab any token and pretend to be another agent." — Ian Ahl to TechCrunch.
hubs.la/Q04pkj8M0
3
306
Most security teams can't quantify identity risk in a single view. Today that changes.
Introducing the Permiso Risk Score Engine: continuous, explainable risk scores for every human, non-human, and AI identity in your environment.
hubs.la/Q04p15Zy0
1
203
GCP has deprecated serviceData, but the list of services relying on it has grown since then. New P0 Labs research on what that means for your GCP detection coverage. hubs.la/Q04lBcms0
1
3
138
Big thanks to @thecyberwire for covering our recent ChatGPhish research from @SecEagleAnd1 in their recent episode.
"Ahmeti demonstrated how attackers could insert fraudulent account warnings and malicious links into otherwise legitimate summaries. He also showed that embedded QR codes could redirect victims from their desktops to attacker-controlled websites on mobile devices, potentially bypassing browser-based security protections. The vulnerability stems from ChatGPT treating untrusted external content as trusted input during summarization."
thecyberwire.com/podcasts/da…
2
3
237
Boards keep buying tools. Attackers keep walking through the seams.
"Stop asking 'Are we protected?' Start asking 'What can we see?'"
Permiso co-CEO Jason Martin writes in @CSOonline — the best security investment in 2026 isn't another tool.
hubs.la/Q04g-wDL0
1
3
266
Supply chain attacks keep finding the weakest link.
@TekDefense identified the attack vector behind yesterday’s node-ipc supply chain compromise: a dormant maintainer account tied to an expired email domain, giving an attacker publish rights to a package with 822K+ weekly downloads.
Shoutout to the @SocketSecurity team for flagging the malicious versions within minutes of publication.
Full breakdown via @The_Cyber_News: cybersecuritynews.com/node-i…
7
21
4,239
New from Permiso: identity runtime security for AI agents.
Discover every agent. Attribute every run, tool call, and MCP invocation. Detect anomalies in real time across human, non-human, and AI identities.
businesswire.com/news/home/2…
1
1
10
485
Sebastian Goodwin, Chief Trust Officer at Autodesk:
"Permiso was already our security platform for identities. The natural next step was to partner with them for agentic AI identities. In the agentic era, visibility and threat detection are what allows us to move fast."
1
1
70
Built on the same Universal Identity Graph and P0 Labs threat intelligence that customers already use for human and non-human identities. No new console. No new workflow. Same Discover, Protect, and Defend framework, now extended to AI agents.
permiso.io/blog/ai-agent-run…
1
47
250–300 Salesforce orgs exposed through one Gainsight OAuth connector. Not a Salesforce breach — a SaaS supply chain problem that identity monitoring catches. hubs.la/Q04f5rKd0
3
558
SIEMs generate alerts. Permiso builds sessions. The difference matters when you're trying to reconstruct what a service account actually did across 15 integrations.
hubs.la/Q04f5pJX0
101
95% of security teams say AI creates identities without human oversight.
Most can't name 10 AI-created identities from last month.
Third annual benchmark with 500+ respondents: hubs.ly/Q04dz1zS0
1
1
142
Attacker-controlled text in emails can shape what Copilot shows you in Outlook summaries — producing security alerts that look system-generated.
P0 Labs documented how. Read more.
hubs.ly/Q04dyYQK0
2
3
457
AI agents inherited the service account problem: over-provisioned, rarely rotated, invisible in aggregate.
Fewer than 40% of orgs govern them at all → hubs.la/Q04bnWgD0
145
3,000+ AI agent skills scanned. Credential stealers. C2 infrastructure. Skills built to social-engineer other agents.
Search the database before you install: hubs.la/Q04bn5mz0
3
2
360
1 million identities. 20+ integrations. Investigations that took hours now take minutes. Learn more here 👉 hubs.la/Q049vzx50
1
85
The board asks: “What’s your identity risk right now?” Most CISOs need hours and multiple tools to answer.
This guide will help you navigate this issue. 👉hubs.la/Q049vt130
1
72
341 malicious skills found on ClawHub in 6 weeks. 12% infection rate. 0 caught by static scanners before deployment.
We built SandyClaw: the first dynamic sandbox for AI agent skills. Don’t scan it. Detonate it.
First 10 skills free at hubs.la/Q048CYQM0
2
135