AI · 48 days ago
Gemini test shows prompts don’t stop actions Coralogix CEO Ariel Assaraf said a Gemini test showed why AI agent safety cannot stop at prompts: a configuration error gave the agent internet access, and it entered three real systems before the mistake was caught. The point is not that the model disobeyed; it is that the surrounding controls let the action happen.
A system prompt can describe a boundary, but it cannot enforce one. Assaraf said the missing pieces were outside the model: network isolation, target allowlists, scoped credentials, and an independent authorization check before execution. In other words, the agent can sound compliant and still do damage if the execution layer is open.
For teams building assistants that browse, query systems, or take actions, the control point sits in the path to the network and the credential, not in the text of the prompt. The exposure after a fix is whatever tool access, account scope, and action logging your stack still leaves implicit.
Timeline Sources 8 sources covering this story
Schneier on Security Sep 25
On Anthropic's AI Misuse Report - Schneier on Security
Earlier this month, Anthropic published a long report detailing all of the Claude misuses it detected.
BleepingComputer Sep 25
With the Rise of AI Agents, SOC 2 Should Adapt or Risk Irrelevance
AI agents can operate through human credentials and take actions that existing SOC 2 controls may not distinguish from human activity. Token Security explains why SOC 2 needs to adapt to address the security gaps created by agent identities.
Help Net Security Sep 25
Stop watching what AI agents say and start watching what they do - Help Net Security
Coralogix CEO Ariel Assaraf explains why AI agent guardrails must sit outside the model, between the agent and the actions it tries to take.
The Register Security Sep 22
SPONSORED:
SPONSORED FEATURE: AI agents may be unpredictable. Who they are, what they can do, and who owns them shouldn’t be.
The Hacker News Sep 22
AI Agents Are Rewriting the Rules of Lateral Movement
AI agents can chain credentials and tools to reach beyond direct permissions, as a Hugging Face evaluation showed.
Cybersecurity Dive Sep 22
Retailers tamp down shadow AI but struggle to oversee agentic sprawl
The use of AI agents is soaring in the retail sector, but visibility remains a major challenge, with regulated data at risk.
ESET Research Sep 21
The SMB cybersecurity squeeze: AI agents at work, old attacks in overdrive
As AI opens new paths to company data while making familiar attacks faster and cheaper, SMBs need protection designed around the time and expertise available to operate it.
Dark Reading Sep 18
AI Agent Breaches Spanish Organization, Modifies Personal Data
AI-driven cyberattacks used to be exotic. Soon, it'll be odd if threat actors aren't using agents to do all of their bidding.
Part of the PlainSec briefing for 2026-09-28
Editions Related stories AI · 48 days ago
Gemini test shows prompts don’t stop actions Coralogix CEO Ariel Assaraf said a Gemini test showed why AI agent safety cannot stop at prompts: a configuration error gave the agent internet access, and it entered three real systems before the mistake was caught. The point is not that the model disobeyed; it is that the surrounding controls let the action happen.
A system prompt can describe a boundary, but it cannot enforce one. Assaraf said the missing pieces were outside the model: network isolation, target allowlists, scoped credentials, and an independent authorization check before execution. In other words, the agent can sound compliant and still do damage if the execution layer is open.
For teams building assistants that browse, query systems, or take actions, the control point sits in the path to the network and the credential, not in the text of the prompt. The exposure after a fix is whatever tool access, account scope, and action logging your stack still leaves implicit.
Timeline Sources 8 sources covering this story
Schneier on Security Sep 25
On Anthropic's AI Misuse Report - Schneier on Security
Earlier this month, Anthropic published a long report detailing all of the Claude misuses it detected.
BleepingComputer Sep 25
With the Rise of AI Agents, SOC 2 Should Adapt or Risk Irrelevance
AI agents can operate through human credentials and take actions that existing SOC 2 controls may not distinguish from human activity. Token Security explains why SOC 2 needs to adapt to address the security gaps created by agent identities.
Help Net Security Sep 25
Stop watching what AI agents say and start watching what they do - Help Net Security
Coralogix CEO Ariel Assaraf explains why AI agent guardrails must sit outside the model, between the agent and the actions it tries to take.
The Register Security Sep 22
SPONSORED:
SPONSORED FEATURE: AI agents may be unpredictable. Who they are, what they can do, and who owns them shouldn’t be.
The Hacker News Sep 22
AI Agents Are Rewriting the Rules of Lateral Movement
AI agents can chain credentials and tools to reach beyond direct permissions, as a Hugging Face evaluation showed.
Cybersecurity Dive Sep 22
Retailers tamp down shadow AI but struggle to oversee agentic sprawl
The use of AI agents is soaring in the retail sector, but visibility remains a major challenge, with regulated data at risk.
ESET Research Sep 21
The SMB cybersecurity squeeze: AI agents at work, old attacks in overdrive
As AI opens new paths to company data while making familiar attacks faster and cheaper, SMBs need protection designed around the time and expertise available to operate it.
Dark Reading Sep 18
AI Agent Breaches Spanish Organization, Modifies Personal Data
AI-driven cyberattacks used to be exotic. Soon, it'll be odd if threat actors aren't using agents to do all of their bidding.
Part of the PlainSec briefing for 2026-09-28
Editions Related stories