Thoughts about this Microsoft advisory?
"Additionally, agentic AI applications introduce novel security risks, such as cross-prompt injection (XPIA), where malicious content embedded in UI elements or documents can override agent instructions, leading to unintended actions like data exfiltration or malware installation."
https://support.microsoft.com/en-us/windows/experimental-agentic-features-a25ede8a-e4c2-4841-85a8-44839191dfb3
Dan Goodin
npub1z3…8qvzu
2025-11-18 18:54:33 UTC
Author Public Key
npub1z3lwfekw80j4ngzg6ky3ks202xr6uwnd4jttxzsd4euc9l55euvq48qvzuSeen on
wss://relay.momostr.pinkPublished at
2025-11-18 18:54:33 UTCEvent JSON
{
"id": "96b73a26d2c95523e92ef4005c7b8e8f1d014607a2aab67b318e5cf6f568ca0b",
"pubkey": "147ee4e6ce3be559a048d5891b414f5187ae3a6dac96b30a0dae7982fe94cf18",
"created_at": 1763492073,
"kind": 1,
"tags": [
[
"proxy",
"https://infosec.exchange/@dangoodin/115572216553053075",
"web"
],
[
"proxy",
"https://infosec.exchange/users/dangoodin/statuses/115572216553053075",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://infosec.exchange/users/dangoodin/statuses/115572216553053075",
"pink.momostr"
],
[
"-"
]
],
"content": "Thoughts about this Microsoft advisory?\n\n\"Additionally, agentic AI applications introduce novel security risks, such as cross-prompt injection (XPIA), where malicious content embedded in UI elements or documents can override agent instructions, leading to unintended actions like data exfiltration or malware installation.\"\n\nhttps://support.microsoft.com/en-us/windows/experimental-agentic-features-a25ede8a-e4c2-4841-85a8-44839191dfb3",
"sig": "e659693acdde098e804eddcf39cdfcc300762e83747f6eec507d6eea00ce35d2cbfe25470a4f991f1d675e9afa32cf79cac2840157de2bc7b9fff79d38667469"
}