Viewing enriched Twitter post
Attacking Vision-Language Agents via Pop-ups As devs begin to use more agents to automate computer tasks, what types of attacks to look out for? This work shows that integrating adversarial pop-ups into existing agent testing environments leads to an attack success rate of 86%. This decreases the agents' task success rate by 47%. They also add that basic defense techniques (e.g., instructing the agent to ignore pop-ups) are ineffective.
{
"score": 0.94,
"scored_at": "2025-08-09T13:46:07.551689",
"import_source": "network_archive_import",
"media": [
{
"type": "photo",
"url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1853810252308774955/media_0.jpg?",
"filename": "media_0.jpg"
},
{
"id": "",
"type": "photo",
"url": null,
"media_url": "https://pbs.twimg.com/media/GboN9VwXMAAsTT4.jpg",
"media_url_https": null,
"display_url": null,
"expanded_url": null
}
],
"reprocessed_at": "2025-08-12T15:25:58.283562",
"reprocessed_reason": "missing_media_array",
"original_structure": "had_both"
}{
"user": {
"created_at": "2015-09-04T12:59:26.000Z",
"default_profile_image": false,
"description": "Building with AI Agents @dair_ai • Prev: Meta AI, Elastic, Galactica LLM, PhD • I also teach how to build with LLMs, RAG & AI Agents ⬇️",
"fast_followers_count": 0,
"favourites_count": 27933,
"followers_count": 216712,
"friends_count": 532,
"has_custom_timelines": true,
"is_translator": false,
"listed_count": 3688,
"location": "",
"media_count": 2656,
"name": "elvis",
"normal_followers_count": 216712,
"possibly_sensitive": false,
"profile_banner_url": "https://pbs.twimg.com/profile_banners/3448284313/1565974901",
"profile_image_url_https": "https://pbs.twimg.com/profile_images/939313677647282181/vZjFWtAn_normal.jpg",
"screen_name": "omarsar0",
"statuses_count": 12439,
"translator_type": "regular",
"url": "https://t.co/JBU5beHQNs",
"verified": true,
"withheld_in_countries": [],
"id_str": "3448284313"
},
"id": "1853810252308774955",
"conversation_id": "1853810252308774955",
"full_text": "Attacking Vision-Language Agents via Pop-ups\n\nAs devs begin to use more agents to automate computer tasks, what types of attacks to look out for?\n\nThis work shows that integrating adversarial pop-ups into existing agent testing environments leads to an attack success rate of 86%.\n\nThis decreases the agents' task success rate by 47%. \n\nThey also add that basic defense techniques (e.g., instructing the agent to ignore pop-ups) are ineffective.",
"reply_count": 5,
"retweet_count": 47,
"favorite_count": 195,
"hashtags": [],
"symbols": [],
"user_mentions": [],
"urls": [],
"media": [
{
"media_url": "https://pbs.twimg.com/media/GboN9VwXMAAsTT4.jpg",
"type": "photo"
}
],
"url": "https://twitter.com/omarsar0/status/1853810252308774955",
"created_at": "2024-11-05T14:43:00.000Z",
"#sort_index": "1853810252308774955",
"view_count": 25353,
"quote_count": 3,
"is_quote_tweet": false,
"is_retweet": false,
"is_pinned": false,
"is_truncated": true,
"startUrl": "https://x.com/omarsar0/status/1853810252308774955"
}