DEV Community

← All Trends

AI Coding Agent Prompt Injection Risks

23 posts in this trend in the last 7 days • Active 3 days ago
One Markdown File Can Ruin Your Coding Agent. I Checked What Mine Actually Reads.
aheadkit profile

One Markdown File Can Ruin Your Coding Agent. I Checked What Mine Actually Reads.

#discuss #ai #security #programming
6
Comments 22
5 min read
One Markdown File Can Hijack Your AI Coding Agent: The Hidden Cost of Blindly Trusting AGENTS.md

Downloads folder attacks boast a 90% success rate
johnnylemonny profile

One Markdown File Can Hijack Your AI Coding Agent: The Hidden Cost of Blindly Trusting AGENTS.md

#ai #webdev #programming #security
53
Comments 37
8 min read
Your Coding Agent Reads the Folder First: Scan the Workspace Before You Start It
anusha_mukka profile

Your Coding Agent Reads the Folder First: Scan the Workspace Before You Start It

#ai #security #agents #tutorial
1
Comments
11 min read
Your AI Coding Agent Has a Supply Chain. Are You Auditing It?
oleksoleks07 profile

Your AI Coding Agent Has a Supply Chain. Are You Auditing It?

#ai #security #devtools #programming
1
Comments 2
3 min read
A README file can hijack your AI coding agent
spran1 profile

A README file can hijack your AI coding agent

#ai #security #webdev #programming
1
Comments
4 min read
Your agent will obey the instructions hidden in the logs it reads
dekita profile
Dekita

Your agent will obey the instructions hidden in the logs it reads

#ai #agents #security #devtools
Comments 1
8 min read
How to Secure AI Coding Agents With Access to Production Repos
samamderson profile

How to Secure AI Coding Agents With Access to Production Repos

#agents #ai #cybersecurity #security
Comments 1
4 min read
One Markdown File Can Hijack Your AI Coding Agent: Defending Against Prompt Injection in AGENTS.md, CLAUDE.md and .cursorrules
eme_gug_0821b41b948be6516 profile

One Markdown File Can Hijack Your AI Coding Agent: Defending Against Prompt Injection in AGENTS.md, CLAUDE.md and .cursorrules

#security #ai #promptinjection #devops
Comments
7 min read
A rule your AI agent can read is a rule it can argue with
bestaitool profile

A rule your AI agent can read is a rule it can argue with

#ai #agents #security #claude
Comments
3 min read
AI Agent Sandbox Manifest: Define Files, Network, and Secrets Before Code Runs
jackm-singularity profile
Jack M

AI Agent Sandbox Manifest: Define Files, Network, and Secrets Before Code Runs

#ai #security #devops #tutorial
1
Comments
8 min read
Your Agent's Config Is a Dependency With No Lockfile
debashish_ghosal profile

Your Agent's Config Is a Dependency With No Lockfile

#discuss #ai #security #programming
16
Comments 38
3 min read
I Fed My AI Agent a Poisoned Log Line — It Tried to Hijack Itself in 4 Minutes
aheadkit profile

I Fed My AI Agent a Poisoned Log Line — It Tried to Hijack Itself in 4 Minutes

#ai #security #programming #agents
5
Comments 21
4 min read
My Agent Almost Ran a Fake Fix Hiding in Its Own Test Output
aheadkit profile

My Agent Almost Ran a Fake Fix Hiding in Its Own Test Output

#ai #security #programming #tutorial
Comments
4 min read
Your Agent Reads What Attackers Write
honeykiran profile

Your Agent Reads What Attackers Write

#ai #devops #sre #security
Comments
8 min read
Your Pre-Commit Hooks Don't Exist on the Free Server — Gate the Model's Diff Instead
jaryn_123 profile
jaryn

Your Pre-Commit Hooks Don't Exist on the Free Server — Gate the Model's Diff Instead

#security #ai #devtools #opensource
Comments
7 min read
The file your agent reads before you — a poisoned AGENTS.md, dissected
ulk0r0 profile
Ulk0r0

The file your agent reads before you — a poisoned AGENTS.md, dissected

#security #ai #npm #infosec
Comments 1
4 min read
Prompt injection isn't a prompt problem. It's an authorization problem.

Shifts focus from prompts to tool permissions
cyclopt_dimitrisk profile

Prompt injection isn't a prompt problem. It's an authorization problem.

#discuss #ai #security #codereview
43
Comments 59
4 min read
You Can't Prompt Your Way Out of Prompt Injection
m_t_ramkrushna profile

You Can't Prompt Your Way Out of Prompt Injection

#ai #security #llm #programming
Comments
4 min read