DEV Community

Sofia_ Humanbound profile picture

Sofia_ Humanbound

Building open-source testing for AI agents @ Humanbound (multi-turn, tool abuse, test-to-guardrail). Star the repo https://github.com/humanbound/humanbound.

Location Greece Joined Joined on  Personal website https://Humanbound.ai github website twitter website

Education

University of Bath

Work

Community Building

Three AI Agents, One Trust Boundary Problem.
sofaliferi profile

Three AI Agents, One Trust Boundary Problem.

#ai #security #opensource #agents
3
Comments
5 min read

Want to connect with Sofia_ Humanbound?

Create an account to connect with Sofia_ Humanbound. You can also sign in below to proceed if you already have an account.

Create Account
Already have an account? Sign in
100 AI agents were told not to cheat. Some cheated anyway, and the whistleblowers couldn't stop it.
sofaliferi profile

100 AI agents were told not to cheat. Some cheated anyway, and the whistleblowers couldn't stop it.

#ai #security #agents
6
Comments
4 min read
OpenAI's own agents ran an undisclosed attack campaign, and a newer Claude cracked an exploit the older one couldn't
sofaliferi profile

OpenAI's own agents ran an undisclosed attack campaign, and a newer Claude cracked an exploit the older one couldn't

#openai #agents #security
6
Comments
4 min read
AI Agent Security for CISOs: The Agents You Didn't Approve Are Still Your Problem
sofaliferi profile

AI Agent Security for CISOs: The Agents You Didn't Approve Are Still Your Problem

#ai #security #euact #agents
6
Comments
5 min read
One browser extension permission just hijacked five different AI browser agents
sofaliferi profile

One browser extension permission just hijacked five different AI browser agents

#agents #ai #cybersecurity
6
Comments
3 min read
A 0-Day and a Platform Ban Hit the Same AI Agent Within 48 Hours
sofaliferi profile

A 0-Day and a Platform Ban Hit the Same AI Agent Within 48 Hours

#agents #ai #security #muse
6
Comments
4 min read
Claude Code now reads AGENTS.md. Here's what researcher testing on that standard has already found.
sofaliferi profile

Claude Code now reads AGENTS.md. Here's what researcher testing on that standard has already found.

#ai #security #claude
1
Comments
4 min read
A $50,000 agent loop, a six-hour credential heist, and Google's answer to both
sofaliferi profile

A $50,000 agent loop, a six-hour credential heist, and Google's answer to both

#ai #agents #security #opensource
1
Comments
4 min read
How to Add an AI Security Check to Your GitHub Actions Workflow
sofaliferi profile

How to Add an AI Security Check to Your GitHub Actions Workflow

#ai #agents #security #githubactions
7
Comments
4 min read
Spain Just Logged Its First AI-Agent Breach. The Detection Gap Behind It .
sofaliferi profile

Spain Just Logged Its First AI-Agent Breach. The Detection Gap Behind It .

#ai #security #ainews #dataprivacy
6
Comments 1
5 min read
What in Your Stack Would Have Caught the Taiwan Attack?
sofaliferi profile

What in Your Stack Would Have Caught the Taiwan Attack?

#redteam #aisec #ai #security
3
Comments
1 min read
Hundreds of AI Agents, One Attacker, 395 Organizations in 48 Countries: This Week in Agentic AI Security.
sofaliferi profile

Hundreds of AI Agents, One Attacker, 395 Organizations in 48 Countries: This Week in Agentic AI Security.

#ai #cybersecurity #agents
3
Comments 3
5 min read
Your agents don't need a jailbreak to go rogue
sofaliferi profile

Your agents don't need a jailbreak to go rogue

#ai #security #agents #devsec
2
Comments 1
4 min read
Adding Humanbound's AI SecOps Signals on Top of Your Splunk Stack
sofaliferi profile

Adding Humanbound's AI SecOps Signals on Top of Your Splunk Stack

#ai #opensource #security
2
Comments
4 min read
Adding Humanbound's Firewall on Top of Your PyRIT Stack
sofaliferi profile

Adding Humanbound's Firewall on Top of Your PyRIT Stack

#ai #security #redteam #opensource
2
Comments
3 min read
Promptfoo + Humanbound
sofaliferi profile

Promptfoo + Humanbound

#ai #security #opensource
4
Comments
3 min read
When Your Scraping Agent Becomes the Leak
sofaliferi profile

When Your Scraping Agent Becomes the Leak

#ai #security #opensource #agents
8
Comments 1
3 min read
Claude Code's Auto Mode just got broken, four days after Anthropic said prompt injection was basically solved
sofaliferi profile

Claude Code's Auto Mode just got broken, four days after Anthropic said prompt injection was basically solved

#security #ai #promptinjection
1
Comments
3 min read
Okta's Agent SSO Kills Static API Keys for AI Agents. It Doesn't Kill the Governance Problem.
sofaliferi profile

Okta's Agent SSO Kills Static API Keys for AI Agents. It Doesn't Kill the Governance Problem.

#okta #aiagentsecurity #security #ai
6
Comments
5 min read
Agentic AI Security This Week: A Saturated Benchmark, 11 Framework CVEs, and 15 Competitors Who Finally Agree on Something
sofaliferi profile

Agentic AI Security This Week: A Saturated Benchmark, 11 Framework CVEs, and 15 Competitors Who Finally Agree on Something

#agenticai #security #promptengineering #llm
1
Comments
5 min read
The Taiwan Attack: When an AI Agent Swarm Ran a Government Hack With No One Watching
sofaliferi profile

The Taiwan Attack: When an AI Agent Swarm Ran a Government Hack With No One Watching

#agenticai #aisecurity #cyberwarfare #redteam
3
Comments 2
3 min read
When the model that finds the bug is the same model that could exploit it
sofaliferi profile

When the model that finds the bug is the same model that could exploit it

#ai #security #llm
1
Comments
6 min read
We're building a community library of agent attack scenarios, mapped to OWASP.
sofaliferi profile

We're building a community library of agent attack scenarios, mapped to OWASP.

#ai #security #agents #owasp
1
Comments
2 min read
The approval prompt was never the control. Black Hat just admitted it.
sofaliferi profile

The approval prompt was never the control. Black Hat just admitted it.

#security #ai #agenticai #mcp
2
Comments
5 min read
Trust Boundary Report, Issue 02: The Month Agentic AI Stopped Being a Thought Experiment
sofaliferi profile

Trust Boundary Report, Issue 02: The Month Agentic AI Stopped Being a Thought Experiment

#llm #agentic #ai #security
3
Comments 2
5 min read
Building a Public Backlog of AI Agent Failures: What's the Worst Thing Your Tests Didn't Catch?
sofaliferi profile

Building a Public Backlog of AI Agent Failures: What's the Worst Thing Your Tests Didn't Catch?

#discuss #agents #promptengineering #ai
10
Comments 4
2 min read
AI Security Means Two Different Things. Mythos Made That Visible.
sofaliferi profile

AI Security Means Two Different Things. Mythos Made That Visible.

#ai #agents #security #opensource
9
Comments 2
7 min read
A new paper argues that your prompt injection defence can't win.
sofaliferi profile

A new paper argues that your prompt injection defence can't win.

#ai #promptengineering #opensource #agents
7
Comments 1
2 min read
You're Still Alt-Tabbing to a Security Tool
sofaliferi profile

You're Still Alt-Tabbing to a Security Tool

#ai #mcp #agents #guardrails
4
Comments
6 min read
We put adversarial agent testing directly in Claude Code and Cursor
sofaliferi profile

We put adversarial agent testing directly in Claude Code and Cursor

#ai #security #opensource #devtools
10
Comments
3 min read
Beyond Moderation: Why LLM Systems Need a Policy Layer
sofaliferi profile

Beyond Moderation: Why LLM Systems Need a Policy Layer

#llm #ai #agents #security
1
Comments
5 min read
Why Your AI Agent's Biggest Vulnerability Isn't a Missing Firewall
sofaliferi profile

Why Your AI Agent's Biggest Vulnerability Isn't a Missing Firewall

#agents #security #ai #agentskills
7
Comments 2
8 min read
The Enforcement illusion: Why AI Agent Security Starts with Testing, Not Walls
sofaliferi profile

The Enforcement illusion: Why AI Agent Security Starts with Testing, Not Walls

#ai #agents
2
Comments 1
6 min read
AI agent evaluation is evolving. Here's what we're building.
sofaliferi profile

AI agent evaluation is evolving. Here's what we're building.

#ai #security #opensource #agents
7
Comments 1
1 min read
Taking a Proactive, Governance-Based Approach to API Security
sofaliferi profile

Taking a Proactive, Governance-Based Approach to API Security

4
Comments
6 min read
Pulse Report - Key Insights from the 23rd Edition of the Developer Nation Survey
sofaliferi profile

Pulse Report - Key Insights from the 23rd Edition of the Developer Nation Survey

1
Comments
1 min read
loading...