Arrakis report: 76% of enterprise agents sit dormant with live credentials. I build agent governors — here's the finding that worries me most
Description
The Arrakis 2026 Autonomous Agents Threat Report reveals that a significant majority of enterprise autonomous agents remain dormant yet retain active credentials, posing security risks. Notably, 76% of agents have not run in 60 days but still have valid credentials and active triggers. Additionally, 24% of dormant agents can be activated externally via unsigned webhooks, and 20% can delegate tasks to other agents with higher permissions, potentially bypassing intended access controls. The report highlights a common operational gap where agent credentials and approvals are not revoked or expired after inactivity, unlike employee access which is promptly removed upon departure.
Reddit Discussion
Arrakis Security published their 2026 Autonomous Agents Threat Report yesterday, from telemetry on thousands of production agents. The numbers:
- 76% of active agents haven't run in 60 days — credentials still work, triggers still fire
- 44% haven't run in 180 days
- 24% of dormant agents can be started by an unsigned webhook from outside the organization
- 20% can pass a task to another agent with different permissions and get back results they aren't allowed to produce themselves
Their CEO's line: when an employee leaves, access is removed the same day. When an agent stops being used, nobody removes anything, because nobody owns it.
Full disclosure: I build in this space. QuietOS is my runtime governor (built on a 4GB Motorola phone, open-source, public proof log — some of you have red-teamed it here). The rules I landed on exist because of exactly this report-shaped problem: consequential actions need a signed human decision, approvals expire instead of living forever (expiry = denied), and licensed installs fade into a safe lock rather than dying armed.
The finding I can't shake is the unsigned webhook. A dormant, credentialed agent that the outside world can wake up is why I'm building the inbound half now — Sentinel: everything coming IN gets inspected before the brain ever sees it, and world content is data, never orders.
Genuine question for people running agents in production: who owns your dormant ones? Does anyone expire agent credentials or approvals, or is it permanent keys all the way down? My sample size is me, and the report says that's the norm — curious if this sub is the exception.
Links cited in this discussion
AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
The report from Arrakis Security, based on telemetry from thousands of production autonomous agents, identifies that 76% of enterprise agents are dormant for over 60 days yet maintain live credentials and active triggers. A substantial portion of these dormant agents (24%) can be externally triggered through unsigned webhooks, creating an attack surface that allows unauthorized activation. Furthermore, 20% of agents have the capability to delegate tasks to other agents with different permissions and receive results beyond their own authorization scope. The CEO of Arrakis emphasizes that while employee access is revoked promptly, dormant agents lack ownership and credential expiration, leading to persistent security exposures. Arrakis offers a behavioral intelligence platform that continuously monitors agent behavior and permissions to detect anomalies and enforce policy adherence.
Potential Impact
The persistence of dormant agents with live credentials and active triggers increases the attack surface within enterprise environments. Unsigned webhooks enabling external activation of dormant agents can be exploited to initiate unauthorized actions. The ability of some agents to delegate tasks and obtain results beyond their permission scope can lead to privilege escalation and unauthorized data access. These factors collectively raise the risk of unauthorized access, data leakage, and potential misuse of autonomous agents in production environments.
Defensive Guidance
No official patch or fix is applicable as this is an operational security issue rather than a software vulnerability. Organizations should implement strict lifecycle management for autonomous agents, including revoking or expiring credentials and approvals for dormant agents. Employing behavioral intelligence solutions, such as those provided by Arrakis, can help monitor agent activity and enforce policies that require signed human decisions for consequential actions. Specifically, disabling or securing unsigned webhooks to prevent unauthorized external activation is critical. Regular audits to identify and remediate dormant agents with live credentials are recommended.
Technical Details
- Source Type
- Subreddit
- cybersecurity
- Reddit Score
- 0
- Discussion Level
- minimal
- Content Source
- reddit_link_post
- Post Type
- link
- Newsworthiness Assessment
- {"score":27,"reasons":["external_link","established_author","very_recent"],"isNewsworthy":true}
- Has External Source
- true
- Trusted Domain
- false
Threat ID: 6adbd30cd73cc49347c1e94f
Added to database: 10/23/2026, 21:35:08 UTC
Last enriched: 10/23/2026, 21:35:11 UTC
Last updated: 10/23/2026, 23:35:04 UTC
Views: 7
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
External Links
Need more coverage?
Upgrade to Pro Console for AI refresh and higher limits.
For incident response and remediation, OffSeq services can help resolve threats faster.
Latest Threats
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.