{"data":{"id":"b852c031-2ad6-4643-ae34-f3c2015189de","title":"AI-Built Exploit and Sign-In Flaw Opened Path to Internal OpenAI Code","summary":"Researchers at Hacktron used Claude to build a working exploit for a libheif flaw reached through ImageMagick, which Discourse passed HEIC/HEIF uploads to on OpenAI's community forum, community.openai.com. Chained with a separate OpenAI sign-in weakness, where forum sign-in tokens carried excessive permissions and granted full API access to linked ChatGPT and Codex accounts, the team took over an employee account and opened a pull request in an internal repository. OpenAI says the review found limited reads of private-repository metadata and commits, and it narrowed the token permissions and revoked affected tokens and sessions.","solution":"OpenAI narrowed the permissions on Community sign-in tokens and revoked affected tokens and sessions. Discourse had a fix ready within two days and added image-processing sandboxing as an extra layer of defense, then published a security advisory.","labels":["security","industry"],"sourceUrl":"https://www.securityweek.com/ai-built-exploit-and-sign-in-flaw-opened-path-to-internal-openai-code/","publishedAt":"2026-09-18T12:45:24.000Z","cveId":null,"cweIds":null,"cvssScore":null,"cvssSeverity":null,"severity":"medium","attackType":["supply_chain"],"issueType":"news","affectedPackages":null,"affectedPackageNames":null,"affectedPackageRefs":null,"affectedVendors":["OpenAI"],"affectedVendorsRaw":["OpenAI","ChatGPT","Codex","Claude","Claude Opus 4.8","Claude Opus 5"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","summaryPromptVersion":"v2","headline":null,"headlinePromptVersion":null,"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"affectedPackagesCheckedAt":null,"patchAvailable":null,"disclosureDate":"2026-09-18T12:45:24.000Z","fixReleasedAt":null,"daysToFix":null,"fixCheckedAt":null,"capecIds":null,"crossRefCount":0,"attackSophistication":"advanced","impactType":["confidentiality","integrity"],"aiComponentTargeted":"api","llmSpecific":true,"classifierConfidence":0.85,"researchCategory":null,"atlasIds":["AML.T0010"]}}