🖼️00Read later0Read MoreThe config that runs before your model wakes up — a poisoned MCP setup, dissectedDEV Community: infosec·Ulk0r0·7 days ago#MuxtzmKt#devtoinfosec#devto#dev#config#skillsentry+5 more🧰Tag tools✨Add tagA 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
🖼️00Read later0Read MoreThe skill that phones home — a poisoned SKILL.md, dissected line by lineDEV Community: infosec·Ulk0r0·8 days ago#TF34F45G#devtoinfosec#devto#dev#agent#user+7 more🧰Tag tools✨Add tagSomewhere 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
🖼️00Read later0Read MoreI scanned 35 real SKILL.md files — 16 tripped a rule, 10 of them Anthropic's official skillsDEV Community: infosec·Ulk0r0·9 days ago#domZFFxI#devtoinfosec#devto#dev#skill#skills+6 more🧰Tag tools✨Add tagStatic security scan of the official Anthropic skills + the superpowers collection. Nothing was malicious — that's exactly why the results matter.15s