Field noteNov 16, 20265 min read1 source

Bitsight published a follow-up on PHP2Shell’s pre-disclosure warning and exploitation timeline

In one minute

Bitsight’s November 11 account reconstructs the warning and patch timeline for PHP2Shell, including an Amazon honeypot observation dated October 17 and public discussion on October 20. For security teams, the central lesson is that vulnerability monitoring can start before CVE-centered alerts—but timestamps and evidence types need careful validation.

What changed: Bitsight adds a follow-up detailing PHP2Shell’s pre-disclosure warning and patch timeline.

What happened

Bitsight describes PHP2Shell as two flaws in PHP request handling: CVE-2026-109618, involving an out-of-memory condition that can leave request data partially initialized before cleanup, and CVE-2026-109619, involving stale request state that may be reused after its memory has been freed. PHP confirmed that both could enable unauthenticated remote code execution on servers handling web requests.

PHP’s security team announced on October 20 that fixes would be released the next day, with 14:00 UTC as the scheduled time. The fixes covered PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35. The announcements for PHP 8.5.12 and 8.4.27 followed the scheduled time, at 14:10 and 14:32 UTC, respectively.

Bitsight says its October 27 review of its threat-intelligence collection found relevant non-repository discussions dated October 20: an ihonker forum thread at 16:00 UTC, an X item at 17:04 UTC and a watchTowr analysis at 18:26 UTC. Bitsight places the public discussion roughly 20 hours before the first patch announcement. The company says it cross-referenced PHP-specific terms with disclosure and exploit terminology and manually reviewed results.

A separate, earlier signal came from Amazon’s MadPot honeypot network. Amazon later reported an exploitation attempt dated October 17, several days before public discussion. Bitsight emphasizes that the attempt’s success was not verified. It also examined GitHub results timestamped October 15–18, but excluded them as evidence of early leak activity: those dates referred to repository creation, while matching text could have been added later. Some indexed content cited the October 21 advisory and patched releases.

Bitsight places PHP’s reach in a supply-chain context, citing its 2025 report, in which PHP ranked seventh among 99 large providers by market share, at 82.1%. Its response recommendations include tracking product terms alongside CVE identifiers, validating the origin of timestamps, distinguishing levels of evidence, mapping intelligence to exposed systems and suppliers, and reviewing logs back to the estimated start of activity.

What the archive adds

New Runtime’s earlier coverage established that PHP2Shell exploitation had been confirmed in the wild, while a separate note reported Bitsight’s finding that roughly one-third of vulnerable PHP hosts patched within a week and end-of-life branches barely moved (confirmed exploitation; patching snapshot). This is a new fact on a known story: Bitsight adds a pre-patch timeline and explains why some apparent earlier GitHub dates were misleading.

Trend signal

This update supports the assumption that PHP2Shell exploitation will keep growing while unpatched internet-facing PHP hosts remain (New Runtime’s analysis). The early attempt reinforces the need to treat the patch window as a period of active risk, though this account does not establish the scale or success of attacks. The signal is about urgency and exposure, not a measured forecast of future incidents.

Why now

Bitsight’s November 11 follow-up brings together its review of October 27 threat-intelligence data and Amazon’s later-reported October 17 honeypot observation. The timeline is useful because it distinguishes the first observed probing from public discussion and the October 21 patches, rather than treating the patch date as the beginning of attacker interest.

Connections to other events

New Runtime’s earlier coverage reported successful in-the-wild PHP2Shell remote code execution observed by Perplexity honeypots, beyond earlier reports of attempts. Bitsight’s account adds a different point in the timeline: Amazon recorded an attempt dated October 17, whose success it could not confirm. Together, the accounts distinguish a pre-disclosure probe from evidence of successful compromise.

Why it matters

For teams responsible for internet-facing PHP, a CVE alert can arrive after attackers have already begun probing. Monitoring component names and technical terms alongside CVE identifiers can surface relevant discussion earlier, while asset and supplier inventories help translate that signal into a list of systems to check.

The timeline also shows why early-warning systems need evidence discipline. A repository’s creation date is not necessarily the date a code sample appeared; a forum post is not proof of a breach; and honeypot activity establishes probing, not successful compromise of production systems. Separating those categories helps teams prioritize without treating noisy or ambiguous indicators as confirmed incidents.

Limits and uncertainty

Amazon’s October 17 observation was an exploitation attempt in honeypot telemetry; its success against live production systems was not confirmed. Bitsight’s reviewed discussion dates show when relevant material appeared in its collection, not necessarily the first time any actor or defender knew about the flaw. The GitHub records it excluded do not establish earlier leak activity because repository creation dates do not date the matching content.

The account does not quantify how many systems were probed, how many were compromised, or how much the early public discussion changed patching outcomes. The cited PHP market-share figure is from Bitsight’s 2025 supply-chain report and does not measure the number of vulnerable PHP deployments. New Runtime has not independently verified the telemetry or timeline.

Skill to build now

For the next critical runtime vulnerability, search both CVE identifiers and the affected product, component, or informal name. Before using a timestamp as evidence of first disclosure, identify what it dates—repository creation, commit, post, or indexed text—and check the surrounding content. Label observations separately as discussion, proof of concept, honeypot probe, or confirmed compromise, then map the strongest signals to exposed assets and review logs from the earliest credible activity date.

Reader takeaway

PHP2Shell’s timeline began before the patches: Amazon later dated a honeypot attempt to October 17, while relevant public discussion surfaced on October 20 and PHP’s fixes arrived October 21. The durable lesson is to monitor technologies as well as CVEs, but never treat a search result’s date—or an attempted exploit—as proof of a leak or successful breach.

Evidence / sources

  1. 1bitsight.com/blog/php2shell-vulnerability-patch-timeline-early-warnings
All Field Notes

Next reads

Read next

Open the graph
  1. 01topicSecurityExplore the security topic hub.
  2. 02archiveField NotesOpen the latest editorial analysis.
  3. 03sourcesSourcesSee the original sources this archive cites.
Who read this page?Machine requests, hidden until opened

Loading the privacy-safe route aggregate...

Open the JSON contract