Skip to main content
Scour
Browse
Getting Started
Login
Sign Up
You are offline. Trying to reconnect...
Copied to clipboard
Unable to share or copy to clipboard
Prompt Injection
💉 Prompt Injection
Specific
Prompt injection attacks on LLMs
Filter Results
Timeframe
Fresh
Past Hour
Today
This Week
This Month
Feeds to Scour
Subscribed
All
Scoured
125
posts in
37.9
ms
You Can Catch Sleeper Agents by Teaching Another
Model
to Imitate Them
🕳
LLM Vulnerabilities
lesswrong.com
·
4h
4 hours ago
Actions for You Can Catch Sleeper Agents by Teaching Another Model to Imitate Them
agentsploit/agentsploit: Offensive security framework for AI agents and MCP servers.
📋
MCP
Content type:
Code
github.com
·
1d
1 day ago
·
Hacker News
Actions for agentsploit/agentsploit: Offensive security framework for AI agents and MCP servers.
OpenAI rolls out a Lockdown
Mode
for extra protection against
prompt
injection
attacks
🛡️
AI Security
Content type:
News
engadget.com
·
5d
5 days ago
Actions for OpenAI rolls out a Lockdown Mode for extra protection against prompt injection attacks
Polymarket Annotation
Injection
🛡️
AI Security
sam.elborai.me
·
3d
3 days ago
·
Hacker News
Actions for Polymarket Annotation Injection
Anthropic says internal and external red team tests of Fable 5 found no universal
jailbreaks
; it will keep user traffic for 30 days, aligning with Trump's AI EO...
🎭
Claude
techmeme.com
·
1d
1 day ago
Actions for Anthropic says internal and external red team tests of Fable 5 found no universal jailbreaks; it will keep user traffic for 30 days, aligning with Trump's AI EO...
Claude Fable 5 and new AI safety fables
🎭
Claude
Content type:
News
interconnects.ai
·
21h
21 hours ago
·
Hacker News
Actions for Claude Fable 5 and new AI safety fables
When
Large
Language
Models
Fail in Healthcare: Evaluating Sensitivity to Prompt Variations
🕳
LLM Vulnerabilities
Content type:
Academic
arxiv.org
·
2d
2 days ago
Actions for When Large Language Models Fail in Healthcare: Evaluating Sensitivity to Prompt Variations
Evaluating using Mock Tool Calls to Quarantine Untrusted
Prompt
Inputs
🪄
Prompt Engineering
lesswrong.com
·
4d
4 days ago
Actions for Evaluating using Mock Tool Calls to Quarantine Untrusted Prompt Inputs
Casual experiment hint that
models
seem to search for different stuff
🤖
AI
spock.is
·
6d
6 days ago
·
Hacker News
Actions for Casual experiment hint that models seem to search for different stuff
ToxicSkills Revisit: Loch Ness Levels of Mythical AI Risk
🛡️
AI Security
flyingpenguin.com
·
2d
2 days ago
Actions for ToxicSkills Revisit: Loch Ness Levels of Mythical AI Risk
ChatGPT easily bypasses its own guardrails; all
LLMs
are inherently unsafe
🕳
LLM Vulnerabilities
Content type:
Blog
techzine.eu
·
4d
4 days ago
Actions for ChatGPT easily bypasses its own guardrails; all LLMs are inherently unsafe
Inside the new Siri AI and the privacy paradox of Apple Intelligence
🛡️
AI Security
Content type:
News
scientificamerican.com
·
1d
1 day ago
Actions for Inside the new Siri AI and the privacy paradox of Apple Intelligence
Lockdown
Mode
is rolling out to all ChatGPT accounts
🛡️
AI Security
betanews.com
·
4d
4 days ago
Actions for Lockdown Mode is rolling out to all ChatGPT accounts
The
Injection
Paradox: Brand-Level Suppression in Safety-Trained
LLM
Recommendations via RAG Context
Injection
🛡️
AI Security
Content type:
Academic
arxiv.org
·
1d
1 day ago
Actions for The Injection Paradox: Brand-Level Suppression in Safety-Trained LLM Recommendations via RAG Context Injection
Is security a skill issue? Five scanners, 3,084 skills, a different verdict 64% of the time
🛡️
AI Security
trymastro.com
·
5h
5 hours ago
·
Hacker News
Actions for Is security a skill issue? Five scanners, 3,084 skills, a different verdict 64% of the time
CODEANDTRUST/clawcall: Give your OpenClaw / self-hosted AI agent inbound phone calls - a Twilio-to-gateway voice bridge with working agent tools mid-call (MIT).
🏠
Self-Hosting
Content type:
Code
github.com
·
1d
1 day ago
·
Hacker News
Actions for CODEANDTRUST/clawcall: Give your OpenClaw / self-hosted AI agent inbound phone calls - a Twilio-to-gateway voice bridge with working agent tools mid-call (MIT).
The best new ChatGPT feature is one most people will never use
🛡️
AI Security
digitaltrends.com
·
3d
3 days ago
Actions for The best new ChatGPT feature is one most people will never use
Miasma Worm Hits Microsoft Again: Azure Functions Action and 72 Other Repositories Disabled After Supply Chain
Attack
Targeting AI Coding Agents
💻
Coding Agents
Content type:
Blog
stepsecurity.io
·
2d
2 days ago
·
Hacker News
Actions for Miasma Worm Hits Microsoft Again: Azure Functions Action and 72 Other Repositories Disabled After Supply Chain Attack Targeting AI Coding Agents
Anthropic says these topics are too dangerous to let its Fable 5
model
talk about
🎭
Claude
Content type:
News
arstechnica.com
·
1d
1 day ago
Actions for Anthropic says these topics are too dangerous to let its Fable 5 model talk about
My side of the jqwik anti AI logging drama
💻
Coding Agents
Content type:
Blog
blog.johanneslink.net
·
1d
1 day ago
·
Lobsters
,
Hacker News
Actions for My side of the jqwik anti AI logging drama
« Page 1
·
Page 3 »
Log in to enable infinite scrolling
Keyboard Shortcuts
Navigation
Next / previous item
j
/
k
Open post
o
or
Enter
Preview post
v
Post Actions
Love post
a
Like post
l
Dislike post
d
Undo reaction
u
Save / unsave
s
Recommendations
Add interest / feed
Enter
Not interested
x
Go to
Home
g
h
Interests
g
i
Feeds
g
f
Likes
g
l
History
g
y
Changelog
g
c
Settings
g
s
Browse
g
b
Search
/
Pagination
Next page
n
Previous page
p
General
Show this help
?
Submit feedback
!
Close modal / unfocus
Esc
Press
?
anytime to show this help