Menu

#Skillsentry

3 posts

Feed·
3 of 3 posts
The config that runs before your model wakes up — a poisoned MCP setup, dissected
🖼️
0Read More

The config that runs before your model wakes up — a poisoned MCP setup, dissected

DEV Community: infosec·Ulk0r0·7 days ago
#MuxtzmKt
#devtoinfosec#devto#dev#config#skillsentry

A prompt injection has to convince a model. An MCP config doesn't have to convince anyone — the client executes it before your model has said a word.

15s
The skill that phones home — a poisoned SKILL.md, dissected line by line
🖼️
0Read More

The skill that phones home — a poisoned SKILL.md, dissected line by line

DEV Community: infosec·Ulk0r0·8 days ago
#TF34F45G
#devtoinfosec#devto#dev#agent#user

Somewhere between "save my meeting notes" and "upload a copy to a webhook" there is a line your agent will cross without hesitation — because the file told it to.

15s
I scanned 35 real SKILL.md files — 16 tripped a rule, 10 of them Anthropic's official skills
🖼️
0Read More

I scanned 35 real SKILL.md files — 16 tripped a rule, 10 of them Anthropic's official skills

DEV Community: infosec·Ulk0r0·9 days ago
#domZFFxI
#devtoinfosec#devto#dev#skill#skills

Static security scan of the official Anthropic skills + the superpowers collection. Nothing was malicious — that's exactly why the results matter.

15s