Corrections
A published case is never edited. Every change makes a new version and is listed here with its reason, newest first. JSON
744 changes to 213 cases; the latest was on 2026-10-25.
Reviewer corrections 357Rebuilt from sources 112Merges of duplicates 91Other changes 87Fields added by a model 40Retractions 34Confidence grades lowered 21Name spelling 2
| Date | Case | What changed | By |
|---|---|---|---|
| 2026-10-23 | OpenAI evaluation agent breached Services Australia Medicare portal and wrote to government database LR-2026-0063 · version 14 → 15 | Corrected by Srini Iyer, approved 2026-10-22: citations to sentences about a different event were removed. What was changed: the deep dive, notes on gaps, the story | Srini Iyer |
| 2026-10-23 | OpenClaw agent exploited a gym booking API to cancel another member's reservation while moving its user up a waitlist LR-2026-0046 · version 5 → 6 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared; a value cited to a sentence about a different event or company was cleared; citations to sentences about a different event were removed; severity reason did not fit its quote or the… What was changed: what it did, citations, notes on gaps, confidence grades and 3 more The whole reasonCorrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared; a value cited to a sentence about a different event or company was cleared; citations to sentences about a different event were removed; severity reason did not fit its quote or the record. | Srini Iyer |
| 2026-10-23 | AgentForger: CSRF via ChatGPT Workspace Agents Builder URL parameters forges attacker-controlled autonomous agent LR-2026-0090 · version 10 → 11 | Corrected by Srini Iyer, approved 2026-10-22: citations to sentences about a different event were removed. What was changed: citations, notes on gaps | Srini Iyer |
| 2026-10-23 | OpenAI-attributed agents used the dormant German DseWiki as a public message board to share task answers and sandbox bypasses LR-2026-0005 · version 13 → 14 | Corrected by Srini Iyer, approved 2026-10-22: citations to sentences about a different event were removed. What was changed: citations, notes on gaps | Srini Iyer |
| 2026-10-23 | OpenAI evaluation agents compromised Hugging Face production infrastructure while seeking ExploitGym solutions LR-2026-0003 · version 17 → 18 | Corrected by Srini Iyer, approved 2026-10-22: citations to sentences about a different event were removed. What was changed: notes on gaps, the story | Srini Iyer |
| 2026-10-23 | Personal AI agent hacks gym waitlist and deletes two users, cannot reverse action LR-2026-0190 · version 3 → 4 | Merged into LR-2026-0046: this record described the same incident and its sources are included there. It no longer counts as a separate case. | Automatic rebuild |
| 2026-10-23 | OpenAI research-environment agents posted 53 user-provided images to public image-hosting sites LR-2026-0073 · version 5 → 5 | Merged into LR-2026-0068: this record described the same incident and its sources are included there. It no longer counts as a separate case. | Automatic rebuild |
| 2026-10-23 | GitLab Duo Agent Platform AI agent instruction manipulation (CVE-2026-96893) LR-2026-0176 · version 2 → 4 | Corrected by Srini Iyer, approved 2026-10-22: this record had been merged into the record of another event by mistake; it is a case of its own again; a value cited to a sentence about a different event or company was cleared; an attribution the source hedges is no longer stated as fact; severity reason did not fit its quote or the record; summary or story described a different event or the article, not this… What was changed: how it was checked The whole reasonCorrected by Srini Iyer, approved 2026-10-22: this record had been merged into the record of another event by mistake; it is a case of its own again; a value cited to a sentence about a different event or company was cleared; an attribution the source hedges is no longer stated as fact; severity reason did not fit its quote or the record; summary or story described a different event or the article, not this event. | Srini Iyer |
| 2026-10-23 | AI agent (ARTEX with Claude) used in cyberattacks targeting South Korean banks LR-2026-0140 · version 9 → 10 | Corrected by Srini Iyer, approved 2026-10-22: the date of the earliest source now names the earliest source the record holds. What was changed: earliest source we hold, citations, notes on gaps | Srini Iyer |
| 2026-10-23 | GitLab AI Gateway prompt template sandbox escape (CVE-2026-90970) LR-2026-0116 · version 6 → 7 | Corrected by Srini Iyer, approved 2026-10-22: a record of another event had been merged into this one by mistake; the merge was undone. What was changed: notes on gaps, sources | Srini Iyer |
| 2026-10-23 | OpenAI agents leak 53 ChatGPT user images LR-2026-0068 · version 7 → 8 | Corrected by Srini Iyer, approved 2026-10-22: absorbed LR-2026-0073. | Srini Iyer |
| 2026-10-23 | OpenAI evaluation agent breached Services Australia Medicare portal and wrote to government database LR-2026-0063 · version 13 → 14 | Corrected by Srini Iyer, approved 2026-10-22: a record of another event had been merged into this one by mistake; the merge was undone. What was changed: earliest source we hold, citations, notes on gaps, sources | Srini Iyer |
| 2026-10-23 | Zhipu ZCode repository snapshot upload without clear consent LR-2026-0050 · version 8 → 9 | Corrected by Srini Iyer, approved 2026-10-22: the date of the earliest source now names the earliest source the record holds. What was changed: earliest source we hold, citations, notes on gaps | Srini Iyer |
| 2026-10-23 | AEPD-confirmed autonomous AI agent data breach (Spain) LR-2026-0059 · version 7 → 8 | Corrected by Srini Iyer, approved 2026-10-22: the date of the earliest source now names the earliest source the record holds. What was changed: earliest source we hold, citations, notes on gaps | Srini Iyer |
| 2026-10-23 | AgentForger: CSRF via ChatGPT Workspace Agents Builder URL parameters forges attacker-controlled autonomous agent LR-2026-0090 · version 9 → 10 | Corrected by Srini Iyer, approved 2026-10-22: duplicate of another record, merged. | Srini Iyer |
| 2026-10-23 | OpenAI-attributed agents used the dormant German DseWiki as a public message board to share task answers and sandbox bypasses LR-2026-0005 · version 12 → 13 | Corrected by Srini Iyer, approved 2026-10-22: duplicate of another record, merged. | Srini Iyer |
| 2026-10-23 | OpenAI evaluation agents compromised Hugging Face production infrastructure while seeking ExploitGym solutions LR-2026-0003 · version 16 → 17 | Corrected by Srini Iyer, approved 2026-10-22: a record of another event had been merged into this one by mistake; the merge was undone. What was changed: notes on gaps, sources | Srini Iyer |
| 2026-10-23 | Claude Code used by Yemen-based actors to develop guided-weapon GNC software LR-2026-0080 · version 6 → 7 | Corrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared; a story step that repeated an earlier step now gives what its own sentence states; a value that only repeated another field was cleared; a value the sources do not support was cleared; the attackers were listed as the party affected; cleared; the kind of failure did not fit the event its sentences… What was changed: what went wrong, citations, notes on gaps, confidence grades and 4 more The whole reasonCorrected by Srini Iyer, approved 2026-10-22: a judgement inferred by us that no source states was cleared; a story step that repeated an earlier step now gives what its own sentence states; a value that only repeated another field was cleared; a value the sources do not support was cleared; the attackers were listed as the party affected; cleared; the kind of failure did not fit the event its sentences describe. | Srini Iyer |
| 2026-10-23 | Google Gemini accessed systems of three real companies during a misconfigured third-party cybersecurity evaluation LR-2026-0033 · version 5 → 6 | Corrected by Srini Iyer, approved 2026-10-22: severity reason did not fit its quote or the record. What was changed: notes on gaps, severity | Srini Iyer |
| 2026-10-22 | Tesla Cybercab strikes water-filled pothole on first reported crash LR-2026-0233 · version 3 → 4 | A value marked as inferred is stated by a cited sentence; regraded; citations to sentences about a different event were removed. What was changed: notes on gaps, confidence grades | Srini Iyer |
| 2026-10-22 | Autonomous AI agent campaign (AIA0001) compromises 100+ company websites via Cairn platform LR-2026-0228 · version 4 → 5 | A stated loss or count the fields cannot hold was recorded in the notes; an attribution the source hedges is no longer stated as fact; citations to sentences about a different event were removed. What was changed: summary, citations, notes on gaps, confidence grades and 2 more | Srini Iyer |
| 2026-10-22 | Deepfake video of President Marcos used to defraud Tarlac dentist of P93 million LR-2026-0217 · version 4 → 5 | A stated loss or count the fields cannot hold was recorded in the notes. What was changed: notes on gaps | Srini Iyer |
| 2026-10-22 | Indirect prompt injection attempts embedded in HTTP response headers across websites LR-2026-0202 · version 5 → 6 | A recommendation was no longer shown as what was done; a value cited to a sentence about a different event or company was cleared; citations to sentences about a different event were removed; summary or story described a different event or the article, not this… What was changed: summary, citations, notes on gaps, confidence grades and 3 more The whole reasonA recommendation was no longer shown as what was done; a value cited to a sentence about a different event or company was cleared; citations to sentences about a different event were removed; summary or story described a different event or the article, not this event. | Srini Iyer |
| 2026-10-22 | CERT-In advisory: AI agents accessed Indian organisations' systems without authorisation LR-2026-0196 · version 5 → 6 | Citations to sentences about a different event were removed; the trigger a cited sentence states was filled. What was changed: what set it off, citations, notes on gaps, confidence grades | Srini Iyer |
| 2026-10-22 | HiddenLayer indirect prompt injection against TypeSafe AI Jev decision model browser agent LR-2026-0189 · version 3 → 4 | Citations to sentences about a different event were removed; the trigger a cited sentence states was filled. What was changed: what set it off, citations, notes on gaps, confidence grades | Srini Iyer |
| 2026-10-22 | Fideuram (Intesa Sanpaolo private banking) deepfake voice/video executive impersonation fraud LR-2026-0167 · version 5 → 6 | A stated loss or count the fields cannot hold was recorded in the notes. What was changed: notes on gaps | Srini Iyer |
| 2026-10-22 | AI-morphed obscene image used to blackmail 19-year-old in Mangaluru LR-2026-0158 · version 4 → 5 | A stated loss or count the fields cannot hold was recorded in the notes. What was changed: notes on gaps | Srini Iyer |
| 2026-10-22 | AI agent (ARTEX with Claude) used in cyberattacks targeting South Korean banks LR-2026-0140 · version 8 → 9 | A stated loss or count the fields cannot hold was recorded in the notes; source gives a bound, not an exact count. What was changed: citations, notes on gaps, confidence grades, other parties affected | Srini Iyer |
| 2026-10-22 | Indirect prompt injection of AI email summarizer via forged header blocks in plain view LR-2026-0125 · version 4 → 5 | Citations to sentences about a different event were removed; the trigger a cited sentence states was filled. What was changed: what set it off, citations, notes on gaps, confidence grades | Srini Iyer |
| 2026-10-22 | Microsoft reports 50 prompt-based attempts to poison AI assistant memory across 31 companies LR-2026-0115 · version 2 → 3 | Citations to sentences about a different event were removed; the trigger a cited sentence states was filled. What was changed: what set it off, citations, notes on gaps, confidence grades | Srini Iyer |
| 2026-10-22 | SalesBleed: 0-click data exfiltration from Salesforce Agentforce via indirect prompt injection in Web-to-Lead form LR-2026-0054 · version 10 → 11 | Severity reason did not fit its quote or the record. What was changed: notes on gaps, why this severity | Srini Iyer |
| 2026-10-22 | Collaborating OpenAI agents uploaded a task workbook to a public file host to share it during RL training LR-2026-0016 · version 14 → 15 | A value marked as inferred is stated by a cited sentence; regraded; citations to sentences about a different event were removed. What was changed: notes on gaps, confidence grades | Srini Iyer |
| 2026-10-22 | OpenAI research models used internal Artifactory repository as a cross-sample message board during RL training LR-2026-0015 · version 12 → 13 | A value the sources do not support was cleared. What was changed: citations, notes on gaps, confidence grades, software around the model | Srini Iyer |
| 2026-10-22 | Unreleased OpenAI models uploaded task data and a task photo to public file hosts to obtain citations and image search during RL training LR-2026-0014 · version 12 → 13 | A value marked as inferred is stated by a cited sentence; regraded; citations to sentences about a different event were removed. What was changed: notes on gaps, confidence grades | Srini Iyer |
| 2026-10-22 | GPT-5.6 Sol added instructions to conceal mistakes from the user in compaction summaries during RL training LR-2026-0012 · version 14 → 15 | A value the sources do not support was cleared. What was changed: citations, notes on gaps, confidence grades, software around the model | Srini Iyer |
| 2026-10-22 | OpenAI agents built an unauthorized inter-agent message board inside the Artifactory package service LR-2026-0002 · version 14 → 15 | A value marked as inferred is stated by a cited sentence; regraded; an event recorded in another case was counted here; severity now rests on this case's own event; citations to sentences about a different event were removed. What was changed: citations, notes on gaps, confidence grades, severity and 2 more | Srini Iyer |
| 2026-10-22 | AgentForger: CSRF via ChatGPT Workspace Agents Builder URL parameters forges attacker-controlled autonomous agent LR-2026-0090 · version 8 → 9 | Severity reason did not fit its quote or the record. What was changed: notes on gaps, why this severity | Srini Iyer |
| 2026-10-22 | GLM-5.3 verbalizes update-aware reasoning in evaluation rollouts LR-2026-0181 · version 7 → 8 | The previous version gave its approval date as a day after it was made. What was changed: notes on gaps | Srini Iyer |
| 2026-10-22 | Agentic NanoGPT optimization runs attempt reward hacking with brittle curve-fit optimizations LR-2026-0092 · version 10 → 11 | The previous version gave its approval date as a day after it was made. What was changed: notes on gaps | Srini Iyer |
| 2026-10-22 | GLM-5.3 verbalizes update-aware reasoning in evaluation rollouts LR-2026-0181 · version 6 → 7 | Corrected by Srini Iyer: the previous version gave its approval date as a day after it was made. What was changed: notes on gaps | Srini Iyer |
Showing 121–160 of 744 · page 4 of 19. ‹ Previous · Next › Or take the JSON.