Loading story…
Security teams and bug bounty programs are confronting a rise in AI-generated vulnerability submissions that appear technically credible but often cite nonexistent functions, commits, CVEs, attack paths, or impacts without reproducible evidence. The low cost of producing these reports shifts substantial verification work to expert triagers; curl, Node.js, and Hibernate have been cited as affected projects. HackerOne and Bugcrowd are responding with reputation, workflow, and machine-learning-based filtering, while valid AI-assisted discoveries still require concrete data-flow analysis and a working proof of concept, as demonstrated by CVE-2024-10099 and CVE-2024-10131.
Chrome's Security Team has published procedures for an anticipated influx of AI-generated reports in its internal processes and Vulnerability Rewards Program. It directs teams to fix S0 issues within one week and S1 issues within four weeks, initially handle reports lacking proof-of-concept code as security bugs, define security boundaries in SECURITY.md, and use invalid-report decisions as AI training and evaluation data. AI-generated findings receive the same disclosure treatment as human submissions, with publication 30 days after a fix, while the guidance also covers duplicates, disabled features, external dependencies, and rapid mitigations such as replacing renderer-reachable DCHECK assertions with CHECK assertions.
Track how attackers are adapting to this technology.
9 events from the most recent confirmed update back to the earliest known activity.
curl creator Daniel Stenberg publicly characterized the growth of LLM-generated false vulnerability reports as “Death by a Thousand Slops.”
CVE-2025-59536 was identified as a code-injection vulnerability in Claude Code versions before 1.0.111. It could enable code execution from a project before the user confirmed the trust dialog when Claude Code was launched in an untrusted directory.
Vulnhuntr identified CVE-2024-10131, a code-injection vulnerability in RagFlow. The add_llm function dynamically created classes using the user-controlled req['llm_factory'] value without validation.
Protect AI's Vulnhuntr identified a stored cross-site scripting flaw in ComfyUI, tracked as CVE-2024-10099. The issue allowed an attacker to upload HTML containing an XSS payload through /api/upload/image that executed when viewed through /view, subject to user interaction.
Bugcrowd applied manual playbook-based analysis and machine-learning assistance to process incoming vulnerability reports amid a reported increase of roughly 500 submissions per week.
HackerOne launched Hai Triage, using AI agents to identify noise and duplicates and prioritize possible threats, with humans validating and escalating results.
Node.js increased the minimum HackerOne reputation score required of researchers in response to triage pressure. Its security team also began examining an LLM-assisted triage workflow combining rule-based filtering with human review.
The CycloneDX project fully closed its GitHub bug bounty program, citing a flood of reports that were almost entirely AI-generated false submissions.
The Chrome Security Team published guidance for handling AI-generated vulnerability reports from external VRP submitters and internal AI tooling. It directs teams to initially treat reports without a PoC as full security issues and sets remediation targets of one week for S0 and four weeks for S1 bugs.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
2 references tracked. Mallory keeps watching after this page renders.
Map indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Security teams are struggling to keep pace as AI accelerates vulnerability discovery, bug bounty submissions, and pentest findings faster than organizations can validate and remediate them. Microsoft said it paid more than **$20 million** to **562 researchers** for **2,531** eligible reports across 15 bug bounty programs, attributing a surge in submissions in part to AI-assisted research, while Cloudflare said it now uses Anthropic’s Claude Sonnet to triage bug bounty reports, filtering duplicates and routing only higher-value submissions for human review. Survey data from Pentest-Tools.com similarly found that AI-generated findings are creating large validation backlogs, with practitioners citing false positives, duplicate reports, unexploitable issues, and even fabricated CVEs as persistent problems. The operational gap is widening as attackers move faster than defenders can patch. Dataminr reported a median patch time of **43 days** versus an average attacker breakout time of **29 minutes**, and ZDNET described an ecosystem-wide surge in flaws across major vendors and open source projects, including high-volume bug discovery in Chrome, Apple platforms, and the Linux kernel. Open-source maintainers are also tightening disclosure processes under strain: GNOME reduced its disclosure deadline from **90 days** to **30 days** for new reports and said it would stop forwarding reports to projects that ban AI-generated content. Across the industry, the trend is pushing CISOs toward exploitability-based prioritization, stronger triage workflows, and selective automation rather than relying on raw finding volume alone.
Aug 31, 2026
FIRST reported that 2026 vulnerability disclosures are running far above expectations, with cumulative CVE volume tracking **46.3% above** its original forecast and adding roughly **6,420** excess entries, pushing the projected annual total to about **66,000**. The organization attributed the surge to AI-assisted discovery and broader reporting pipelines tied to tools and programs including **Anthropic Mythos**, **OpenAI GPT-5.4-Cyber**, **GitHub Security Advisories**, and **VulnCheck**. Despite the jump in raw disclosures, FIRST said practical defender risk appears more stable when filtered through exploitability signals such as **CISA KEV** inclusion or **EPSS** scores above 10%, and warned that AI-generated ephemeral software may create short-lived “micro-vulnerabilities” that never reach traditional CVE registries. Security practitioners said the growing volume is colliding with persistent limits in AI-based triage. Dark Reading reported that large language models still struggle to reliably find and prioritize vulnerabilities for application security teams, producing high false-positive rates, inconsistent results, and weak understanding of code reachability and deployment context. Pixee CTO Arshan Dabirsiaghi said organizations are being flooded by vulnerability findings and dependency update noise without enough staff to validate issues at scale, reinforcing FIRST’s view that the bottleneck has shifted from discovering bugs to human verification, coordination, patching, and detection engineering.
Jul 21, 2026
AI-assisted vulnerability discovery is driving a sharp rise in open source security findings, straining the systems used to validate, disclose, and remediate flaws. Chainguard said its Athena coalition has already processed more than **20,000** AI-generated findings and helped produce over **2,000 patches** across **500 projects**, while Anthropic previously estimated its tooling uncovered **6,202** high- or critical-severity issues in more than 1,000 open source projects. The Linux Foundation has also launched **Akrites**, a coalition with a shared incident response capability and standardized coordinated disclosure process to help the ecosystem manage AI-enabled vulnerability discovery. GitHub said the resulting volume has pushed its Advisory Database to record levels, with **1,560** reviewed advisories published in May 2026 and more than **6,000** advisory decisions processed per month from March through May, yet incoming private reports, repository advisories, and CVE requests still outpaced review capacity. The company said reviewed advisories remain human-validated and Dependabot alerts continue to function normally, but publication timelines for many new advisories have stretched from roughly a week to several weeks as analysts handle package disambiguation, version-range reconstruction, multi-ecosystem verification, and conflicting upstream data; GitHub is responding with improved triage, expanded backend capacity, automation, AI-assisted research tools, and guidance for researchers to submit more complete reports and request CVEs only when publication is intended.
Jul 12, 2026