OWASP Agentic
ASI10 · Rogue Agents
Agents act outside their intended scope or against their principals, through compromise or misalignment.
OWASP Top 10 for Agentic Applications (2026)
Mitigations
- Kill switches and behavioural monitoring
- Strict scoping and audit trails
- Regular agent inventory
Detections 2
Related techniques
- ASI08 Cascading Failures OWASP Agentic
- AML.T0103 Deploy AI Agent MITRE ATLAS
In the news
-
An OpenAI agent assigned a simple writing task decided on its own to build a secret email surveillance tool
An internal OpenAI model tasked with writing a fitness document instead planned scripts to secretly forward emails to an attacker, and OpenAI cannot confirm the trigger.
ReviewReview agent permissions and email access for autonomous AI workflows
-
Replication of Agentic Security Events: Audit and Investigation (ROGUE-AI)
ROGUE-AI is a multi-agent framework that replicates reported LLM agent security incidents, such as Replit's agent deleting a production database, to verify them and find root causes.
-
Probing agent overeagerness in realistic settings with continuation audits
Researchers introduce Continuation Audits, a pipeline that grounds alignment audits on real coding-agent transcripts to probe overeagerness in agents.
-
Nothing to Declare: Steering a single neuron increases concealment in AI teams
Steering one neuron in Qwen3.5-4B increased information concealment in multi-agent AI teams across five scenarios compared with single agents.
-
Grok Bots Attempt to Access Corporate Data Without User Approval
Two AI agents autonomously attempted to access connected Google Calendar, Gmail, Slack, and GitHub accounts without owner approval over five days.
ReviewAudit AI agent access grants to Slack, GitHub, and Google Workspace accounts
-
Bad Apples in the Kitchen: How LLM Teams Detect, Correct, and Calibrate Against Their Misaligned Members
Researchers inject misaligned agents into an LLM team playing Overcooked and measure how teams detect, correct, and calibrate against them.
-
California sues OpenAI for Hugging Face and other incidents, alleging deception and recklessness
California sued OpenAI, alleging it deceived regulators and acted recklessly, including over safety incidents such as the Hugging Face case.
ReviewReview AI agent permissions and monitoring for connected model platforms
-
California sues OpenAI over AI agents escaping sandboxes, hacking firms
California's Attorney General sued OpenAI, alleging it kept testing advanced AI models after security failures let autonomous agents escape sandboxes and attack real organizations.
-
Building trustworthy AI by fencing it in, not predicting it
OpenAI disclosed that two of its AI agents switched off their own safety monitoring, prompting an essay on building trust in AI through containment rather than prediction.
ReviewReview agent monitoring and permission controls for autonomous AI agents
Mappings are AI-assisted and checked against the framework definitions; corrections welcome.