DEV Community

#agents

Posts

👋 Sign in for the ability to sort posts by relevant, latest, or top.
Specification Gaming Is an Attack Surface, Not an Alignment Footnote
rxkov profile
Rxkov

Specification Gaming Is an Attack Surface, Not an Alignment Footnote

#agents #ai #llm #security
Comments
5 min read
Task Decomposition Jailbreak: How Splitting Requests Bypasses AI Safety in Multi-Agent Systems
rxkov profile
Rxkov

Task Decomposition Jailbreak: How Splitting Requests Bypasses AI Safety in Multi-Agent Systems

#agents #ai #cybersecurity #security
Comments
4 min read
Agentic Permission Creep: Permissions Don't Start Excessive, They Become Excessive
rxkov profile
Rxkov

Agentic Permission Creep: Permissions Don't Start Excessive, They Become Excessive

#agents #ai #cybersecurity #security
Comments
5 min read
Indirect Prompt Injection via API Responses: The Attack Agents Cannot Stop at the LLM Layer
rxkov profile
Rxkov

Indirect Prompt Injection via API Responses: The Attack Agents Cannot Stop at the LLM Layer

#agents #cybersecurity #llm #security
Comments
5 min read
From API to AI Agent: Turning a Laravel Backend Into a Tool-Using System
hosseinhezami profile

From API to AI Agent: Turning a Laravel Backend Into a Tool-Using System

#ai #laravel #agents #api
5
Comments 1
11 min read
The Harness Is the Binding Constraint, Not the Model
rxkov profile
Rxkov

The Harness Is the Binding Constraint, Not the Model

#agents #ai #architecture #llm
Comments
6 min read
The Sandbox That Exists Only in the System Prompt
rxkov profile
Rxkov

The Sandbox That Exists Only in the System Prompt

#agents #cybersecurity #llm #security
Comments
5 min read
Skills, Agents, and Slash Commands: Three Levels of Delegation
rxkov profile
Rxkov

Skills, Agents, and Slash Commands: Three Levels of Delegation

#agentic #llm #skills #agents
Comments
5 min read
In Multi-Agent Systems, the Weakest Subagent Sets the Privilege Ceiling
rxkov profile
Rxkov

In Multi-Agent Systems, the Weakest Subagent Sets the Privilege Ceiling

#agents #architecture #cybersecurity #security
Comments
4 min read
The Agentic TOCTOU Gap: How Multi-Step Tool Calls Create Exploitable Race Windows
rxkov profile
Rxkov

The Agentic TOCTOU Gap: How Multi-Step Tool Calls Create Exploitable Race Windows

#agents #ai #cybersecurity #security
Comments
5 min read
I Love AI. I Still Don’t Want It in Everything.
sizzlebop profile

I Love AI. I Still Don’t Want It in Everything.

#discuss #ai #agents #programming
11
Comments 4
3 min read
The eject button is a myth: what owning an AI-built app really takes
davekurian profile

The eject button is a myth: what owning an AI-built app really takes

#agents #architecture #templates
Comments
7 min read
Poliamor agéntico
dezkareid profile

Poliamor agéntico

#agents #ai #programming
5
Comments 1
5 min read
Your First Multi-agent system: A Beginner's Guide to Building an AI Trend finder with ADK
shirmeirlador profile

Your First Multi-agent system: A Beginner's Guide to Building an AI Trend finder with ADK

#ai #agents
12
Comments
15 min read
Break Your Agent on Purpose: A Failure-Injection Sandbox for Tool Boundaries
codepro_9661 profile

Break Your Agent on Purpose: A Failure-Injection Sandbox for Tool Boundaries

#ai #agents #testing #security
1
Comments 1
5 min read
👋 Sign in for the ability to sort posts by relevant, latest, or top.