--- title: Exploit brokers pay $500,000 for a WordPress RCE. I... description: A security researcher at Searchlight Cyber used GPT-5.6 Sol Ultra with a multi-agent prompt (adapted from OpenAI's Cycle Double Cover conjecture prompt) to... canonical: https://daily.dev/posts/exploit-brokers-pay-500-000-for-a-wordpress-rce-i-found-one-with-gpt5-6-sol-ultra-and-25-search-lgmv26dun twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Exploit brokers pay $500,000 for a WordPress RCE. I found one with GPT5.6 Sol Ultra and $25 › Searchlight Cyber | daily.dev og:description: A security researcher at Searchlight Cyber used GPT-5.6 Sol Ultra with a multi-agent prompt (adapted from OpenAI's Cycle Double Cover conjecture prompt) to... og:url: https://daily.dev/posts/exploit-brokers-pay-500-000-for-a-wordpress-rce-i-found-one-with-gpt5-6-sol-ultra-and-25-search-lgmv26dun og:image: https://api.daily.dev/og/posts/LGMv26dUN.png og:image:alt: Exploit brokers pay $500,000 for a WordPress RCE. I found one with GPT5.6 Sol Ultra and $25 › Searchlight Cyber og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Exploit brokers pay $500,000 for a WordPress RCE. I found one with GPT5.6 Sol Ultra and $25 › Searchlight Cyber **[Hacker News](https://daily.dev/sources/hn)** · 36 min read · 2 upvotes · 0 comments ## Summary A security researcher at Searchlight Cyber used GPT-5.6 Sol Ultra with a multi-agent prompt (adapted from OpenAI's Cycle Double Cover conjecture prompt) to discover a pre-authentication RCE exploit chain in WordPress for roughly $25. The chain begins with a desync bug in the WordPress Batch API that bypasses parameter sanitization, enabling a pre-auth SQL injection via the author__not_in parameter. From there, the exploit abuses WordPress's in-memory post cache, oEmbed caching, customize_changeset post types, cycle detection logic, and dynamic action hooks to temporarily assume administrator identity and replay the batch request with elevated privileges — ultimately creating a new admin account and uploading a backdoor plugin. The full exploit was produced in ~10 hours. The researcher argues this represents a significant capability leap for AI in security research, with the model chaining multiple obscure gadgets across a large codebase with precision previously considered uniquely human. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How does the WordPress batch REST API validation bypass vulnerability work? The WordPress batch API (introduced in WordPress 5.6) validates and sanitizes all batched requests in one loop, then executes them in a second loop, using two arrays (matches and validation) intended to stay index-aligned. If an early request is malformed, its validation entry updates but the matches array does not, shifting matches out of sync with validation by one position. This lets an attacker validate one request but execute a different, unsanitized endpoint handler, bypassing parameter sanitization entirely. _Teams running WordPress can track exploit chain writeups like this on daily.dev to patch before real attackers act._ ### How was the WordPress batch API SQLi escalated from read-only to full RCE? The pre-auth SQL injection in the GET /wp/v2/posts author__not_in parameter let an attacker fabricate fake in-memory posts via UNION-based injection. Chaining oEmbed cache poisoning, a customize_changeset settings row, and a post-parent cycle-detection gadget that triggers the parse_request hook, an attacker could temporarily assume an administrator's identity, create a new admin account, and upload a backdoor plugin for code execution. _Understanding chained exploit primitives like cache poisoning and hook abuse helps developers harden similar architectures, tracked on daily.dev._ ### How much did it cost to find a pre-auth WordPress RCE using GPT5.6 Sol Ultra? The full exploit chain, from initial SQL injection discovery to complete RCE, was produced using roughly $25 worth of usage on a $200 monthly AI subscription plan (about 50% of a weekly usage allocation), taking just over 10 hours total including four hours of autonomous multi-agent research to escalate the bug to RCE. _Developers weighing AI-assisted security research costs against traditional pentesting can follow cases like this on daily.dev._ ## Similar posts on daily.dev - [wp2shell: WordPress Patches a Pre-Auth RCE That Needed No Plugins](https://daily.dev/posts/wp2shell-wordpress-patches-a-pre-auth-rce-that-needed-no-plugins-pmh9adxqu) · Latest Hacking News · 2 upvotes · 0 comments --- Tags: [#php](https://daily.dev/tags/php), [#sql](https://daily.dev/tags/sql), [#wordpress](https://daily.dev/tags/wordpress) [View this post on daily.dev](https://daily.dev/posts/exploit-brokers-pay-500-000-for-a-wordpress-rce-i-found-one-with-gpt5-6-sol-ultra-and-25-search-lgmv26dun) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Exploit brokers pay $500,000 for a WordPress RCE. I found one with GPT5.6 Sol Ultra and $25 › Searchlight Cyber","url":"https://daily.dev/posts/exploit-brokers-pay-500-000-for-a-wordpress-rce-i-found-one-with-gpt5-6-sol-ultra-and-25-search-lgmv26dun","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/exploit-brokers-pay-500-000-for-a-wordpress-rce-i-found-one-with-gpt5-6-sol-ultra-and-25-search-lgmv26dun"},"datePublished":"2026-07-20T09:37:40.240Z","dateModified":"2026-09-13T21:18:38.221Z","description":"A security researcher at Searchlight Cyber used GPT-5.6 Sol Ultra with a multi-agent prompt (adapted from OpenAI's Cycle Double Cover conjecture prompt) to...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/4e37ca3fd0e8bf41487261e924dce85c?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/4e37ca3fd0e8bf41487261e924dce85c?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Hacker News","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Hacker News","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/hn","url":"https://daily.dev/sources/hn"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/exploit-brokers-pay-500-000-for-a-wordpress-rce-i-found-one-with-gpt5-6-sol-ultra-and-25-search-lgmv26dun","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"php,sql,wordpress","timeRequired":"PT36M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Hacker News","item":"https://daily.dev/sources/hn"},{"@type":"ListItem","position":3,"name":"Exploit brokers pay $500,000 for a WordPress RCE. I found one with GPT5.6 Sol Ultra and $25 › Searchlight Cyber"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/exploit-brokers-pay-500-000-for-a-wordpress-rce-i-found-one-with-gpt5-6-sol-ultra-and-25-search-lgmv26dun#faq","mainEntity":[{"@type":"Question","name":"How does the WordPress batch REST API validation bypass vulnerability work?","acceptedAnswer":{"@type":"Answer","text":"The WordPress batch API (introduced in WordPress 5.6) validates and sanitizes all batched requests in one loop, then executes them in a second loop, using two arrays (matches and validation) intended to stay index-aligned. If an early request is malformed, its validation entry updates but the matches array does not, shifting matches out of sync with validation by one position. This lets an attacker validate one request but execute a different, unsanitized endpoint handler, bypassing parameter sanitization entirely. Teams running WordPress can track exploit chain writeups like this on daily.dev to patch before real attackers act."}},{"@type":"Question","name":"How was the WordPress batch API SQLi escalated from read-only to full RCE?","acceptedAnswer":{"@type":"Answer","text":"The pre-auth SQL injection in the GET /wp/v2/posts author__not_in parameter let an attacker fabricate fake in-memory posts via UNION-based injection. Chaining oEmbed cache poisoning, a customize_changeset settings row, and a post-parent cycle-detection gadget that triggers the parse_request hook, an attacker could temporarily assume an administrator's identity, create a new admin account, and upload a backdoor plugin for code execution. Understanding chained exploit primitives like cache poisoning and hook abuse helps developers harden similar architectures, tracked on daily.dev."}},{"@type":"Question","name":"How much did it cost to find a pre-auth WordPress RCE using GPT5.6 Sol Ultra?","acceptedAnswer":{"@type":"Answer","text":"The full exploit chain, from initial SQL injection discovery to complete RCE, was produced using roughly $25 worth of usage on a $200 monthly AI subscription plan (about 50% of a weekly usage allocation), taking just over 10 hours total including four hours of autonomous multi-agent research to escalate the bug to RCE. Developers weighing AI-assisted security research costs against traditional pentesting can follow cases like this on daily.dev."}}]} ```