CrowdStrike introduces agentic identity provider, parallel SOC investigations and open-source package blocking at Fal.Con 2026

The identity offering, called CrowdStrike Agentic Identity Provider, or Agentic IdP, addresses the step before continuous authorization, according to the company. Continuous Identity, a real-time authorization model CrowdStrike introduced in June and built on technology from its $740 million acquisition of SGNL Inc., decides whether an agent may perform a specific action at a given moment. Agentic IdP establishes what the agent is in the first place.

"Continuous Identity modernized identity security for the agentic era, but you cannot continuously authorize an identity you were never able to establish, and traditional identity providers break the moment an agent acts on its own," said Scott Kriz, general manager of Continuous Identity at CrowdStrike. "Agentic IdP is the identity provider for AI agents."

Companies typically stand in for agent identity today by using service accounts, API keys and workload identities, none of which were designed for software that takes autonomous action on a person's behalf and delegates work to sub-agents. With Agentic IdP, registration flows through Falcon Guardian, an agent discovery and enforcement product CrowdStrike launched earlier at the conference. Guardian finds agents across the enterprise, and Agentic IdP registers each one as it comes online under a single directory. Each agent receives a cryptographically verifiable identity that CrowdStrike said cannot be spoofed or shared. Credentials are not handed to the agent; instead, Agentic IdP brokers tokens scoped to minimum access for the task and the time required, an approach meant to eliminate standing privilege. Every action an agent takes is bound to the human or workload it is acting for, leaving an accountable party.

The second announcement rebuilds how investigations run in Falcon. Charlotte AI now dispatches agents to endpoint, identity, SaaS, cloud and network in parallel rather than in sequence, working a single investigation across all of them at once. CrowdStrike said the change turns work that took hours into minutes. The company said first-generation SOC AI tools send one agent per alert and process alerts sequentially, returning a fragment from one domain instead of a verdict. "AI agents in the SOC are table stakes. Agents working together across every domain, on the same investigation, that's the new standard," said Michael Sentonas, president of CrowdStrike. "CrowdStrike's architecture and expert validation make this possible, and confidently answer the question every CISO is asking: how do I trust what my agents found, and how do I know it's right?"

The parallel agents are held together by a shared context layer, which CrowdStrike described as persistent memory spanning every agent, investigation and tenant. It sits on the Enterprise Graph data layer introduced at last year's Fal.Con. What one agent establishes becomes available to the rest, removing handoffs between them. CrowdStrike said the platform generates close to four trillion events a day across the five domains, and decisions made by its analysts in managed detection and response and incident response engagements are fed back to the agents. Investigations now also include attacks on enterprise AI systems, such as model abuse, prompt injection and data exfiltration through AI assistants; agents return verdicts with staged responses and show the reasoning behind them.

The third feature blocks malicious open-source packages at the endpoint before their embedded code can run, addressing the growing risk of code from unvetted third-party libraries. CrowdStrike did not provide further technical details in the release.

Analysts following the event highlighted the identity shift. Krista Case, principal analyst and practice lead for cyber resilience and security at theCUBE Research, said the continuous-authorization capability was one of the day-two announcements that stood out most. "It's looking at continuous authorization, having that continuous approach, as AI agents are continuously accessing data and taking actions," Case said. "CrowdStrike is scoping access to specific tasks with short-lived authorization that can be revoked when the task ends." She said the model requires understanding how behavior evolves over time and having safeguards to prevent agents from going off track, whether through malicious intent or drift.

CrowdStrike researcher Adam Meyers told the Fal.Con audience that AI agents are generating roughly 250 times more detections than humans. The company's threat research published this week found the average eCrime breakout time has fallen to 29 minutes, with the fastest intrusion clocked at 27 seconds.