What are the bug classes, what was demonstrated, and what is still open?

Fastly: 'most deployments may be safe from RCE. Do not assume this holds true for you.' Imperva: five demonstrated techniques, three via session upload progress; about 67 KB of memory disclosed per attack.

The fix replaces bailout with a delayed error and an UnwindExit at a safe point, adds a hard limit that _exits, and resets the globals. See the code page.

Two researcher parties, no third-party reproduction in the vault. The demo videos were not downloaded.

The Hacker News says the same of the public code. Fastly: working RCE needs knowledge of the running binary's layout on x86_64.

A position, from a vendor; consistent with the Fediverse report of a third bug.

GHSA-w8hw (CVE-2026-109618): 'no complete FPM exploit has been demonstrated for this SAPI'. The Foundation blog: RCE 'was also confirmed on PHP-FPM through application-level code paths'. GHSA-6wmf (CVE-2026-109619): session lifecycle RCE 'validated against mod_php and FPM'. The accounts may refer to different CVEs.

The CVE records list only 8.2 to 8.5 with defaultStatus 'unaffected'. php.net published source backports for 7.4, 8.0 and 8.1, which implies they are affected. A scanner that trusts the CVE record will miss 7.4 to 8.1.

Swiss BACS: 'Even a web application firewall alone is not sufficient.' Cloudflare says its rules blocked the traffic it saw and that attackers mutated payloads to get past them.

The PHP Foundation: 'a wrong CVSS score of 10.0 (the actual score is 9.2)'. Daubois traces it to a social post relayed by Phoenix Security and to watchTowr via Cyber Daily.

Per a Matrix channel, Canonical asked PHP and 'was told that it's unrelated and not a real issue'. Second-hand. Nothing from PHP on the list at capture.

The hash tweet is dated 5 August 2026 on X; its pre-image has not been published, so what it commits to cannot be checked.

tomash-areion is not named in the PHP Foundation's thanks and appears nowhere else on this beat.

Two parties agree. The Foundation calls the delay 'a structural failure'. Cloudflare lists Quarkslab among the 2026 disclosers; nobody else does.

All three research parties say AI was used. Calif: 'No special harness, or access to models that weren't generally available to everyone were used'. Calif does not name a model; Wiz names Gemini 4 Argon, GPT-5.6 Sol and Wiz Atlas; OpenAI's blog reportedly names GPT-5.6 Sol but is blocked.

The GitHub advisories credit amyb-oai among four reporters. OpenAI's own post credits Calif's discovery 'using GPT-5.6 Sol' and Wiz; an X Community Note disputes OpenAI's role.

Wiz's X post says 'Wiz Research uncovers' with no credit to Calif or OpenAI; the blog thanks both.

Wiz's own wording is 'among the most severe vulnerabilities discovered in PHP, and arguably in software history'. OpenAI: 'one of the most impactful vulnerabilities in recent memory'.

The Foundation says the OOM signs were found while fixing one of the audit's findings, not in the audit report itself.

Unverified by anyone else.

Amazon reports an exploit attempt against a honeypot on 17 October, four days before disclosure, from an actor whose source of exploit information is unknown.

The php.net ChangeLog still dates all four releases 22 October. Wiz's X thread still says 22 October.

Primary party. The Foundation does not say who or where.

Contradicts the Foundation's 'no technical details leaked' and The Hacker News's 'a proof-of-concept exploit leaked'. No copy of the post, screenshot or video is in the vault.

Calif's own page does not mention a leak or a deleted video; it embeds two demo videos. Whether one of them is the deleted video is unknown.

Dan Lorenc (via The Register): 'I made the call to get these out early when exploits hit in the wild'; several unnamed vendors had reported exploits; Chainguard's honeypots trapped IPs; a researcher warned the company on the Friday and its team then found the bugs independently with an AI harness. No evidence published. The PHP security team found no functional exploit by release time.