Visa aimed Anthropic’s Claude Mythos at the infrastructure behind billions of each day transactions, a community that spans greater than 200 nations and territories, strikes cash in roughly 160 currencies, and connects practically 5 billion fee credentials to greater than 175 million service provider areas.
The mannequin stitched minor weaknesses deep in the stack into working exploit chains that might historically have surfaced solely late in penetration testing. Rajat Taneja, Visa’s president of know-how, walked the VB Transform 2026 viewers via what got here subsequent, together with why Visa launched the harness that ruled the total hunt as open supply and why the firm deserted conventional remediation metrics for a measurement its staff invented.
Taneja has run know-how technique, product engineering, and world infrastructure at Visa since 2019, after becoming a member of the firm in 2013 from Digital Arts, the place he served as CTO following 15 years at Microsoft. He co-authored, with Visa chief information safety officer Subra Kumaraswamy, the June 10 blog post saying the launch of the Visa Vulnerability Agentic Harness on GitHub as a reference implementation that any safety staff can examine, adapt, and prolong. Visa additionally printed a technical white paper detailing the structure, classes realized, and 12 non-negotiable architectural practices for essential infrastructure.
Belief constructed on pessimism and paranoia
Taneja led with the arithmetic that makes Visa a goal value defending obsessively. Belief at the scale of worldwide funds will get engineered via what he referred to as pessimism and paranoia, by assuming failure and designing round it before failure arrives. The community has been hardened over a few years via zero-trust structure, layered defenses, and extremely automated safety operations constructed for the scale and reliability world funds demand.
So when Anthropic invited the organizations behind essential software program to take a look at Mythos beneath Undertaking Glasswing, Visa stated sure. Glasswing members collectively recognized greater than 10,000 high- or critical-severity vulnerabilities in the first month of testing throughout software program underpinning essential methods industry-wide, in accordance to Anthropic. Anthropic’s personal conclusion positioned the bottleneck after discovery, in verification, disclosure, and patching pace. Visa joined to take a look at many years of hardening at AI pace and study the place superior fashions might push its defenses additional.
What Mythos confirmed at Visa
Inside Visa’s setting, Mythos demonstrated system-wide, context-aware evaluation, surfacing vulnerabilities buried deep in the stack and flagging points that develop extra severe when chained collectively, with findings clear sufficient that engineering groups might act on them with out wading via noise. Some findings carried essential severity scores, and Visa credit its zero-trust controls, community segmentation, and layered safeguards with breaking the chain before any external actor might have acted.
That affirmation mattered, Taneja stated, however the epiphany that adopted mattered extra. “In a world of agentic assaults, protection additionally has to be agentic,” he stated. Even at an organization that has invested many years in defense-in-depth, the mannequin revealed assumptions the staff had been working beneath that wanted rethinking. Conventional SAST instruments hold their place as a primary move towards recognized vulnerability patterns, Visa’s white paper notes, however sample matching alone can’t observe an adversary who causes via logic, information circulate, and the exploit chains that dwell between the signatures.
A harness, not a scanner
Visa’s response was not one other monolithic scanner. The staff constructed the Visa Vulnerability Agentic Harness, now in its fifth technology, as a ruled pipeline that directs frontier AI fashions via structured safety duties whereas imposing deterministic controls, coverage gates, and human oversight at each stage. Taneja walked via the design philosophy. The harness operates throughout 4 phases and eleven phases, from code ingestion and menace modeling via deep-dive verification, exploit chain synthesis, and at last remediation and repair validation.
Three design decisions drive discovering high quality, per the venture’s personal documentation. Risk modeling runs before evaluation to focus on the assault floor quite than scanning the whole lot blindly, multi-agent deterministic voting requires convergence throughout unbiased reasoning chains before a discovering advances, and structured triage artifacts compress the lifecycle from discovery to a end result builders can truly ship. The payoff is a pipeline that runs sizzling by default. A plain scan in the shipped profile runs all eleven phases and edits supply information in the goal repository in repair mode, making use of candidate patches until the operator stops it at detection.
The harness is multi-model by design. An LLM abstraction layer lets Visa swap or mix suppliers with out altering the management aircraft, and the open-source model works with Anthropic Claude, OpenAI-compatible fashions, or a mixture. The repo’s documentation is candid about the exception. Making use of a repair requires the file-editing instruments that solely the Anthropic backends expose, so the remediation and validation phases at present require Anthropic fashions for full performance, and an OpenAI-compatible mannequin in these roles is restricted to report-only output. VentureBeat’s Q2 2026 Pulse analysis, introduced earlier at the convention, reinforces why that supplier flexibility issues. Amongst the enterprises surveyed, 82% rely on provider-native controls as their primary security layer, and 59% plan to undertake or change agent safety tooling inside the 12 months. The controls enterprises adopted final 12 months are already turning into the controls they plan to exchange.
Imply Time to Adapt replaces legacy metrics
Discovering vulnerabilities is not the laborious half, Taneja argued. The true problem is how rapidly a staff can verify a problem is really exploitable, repair it, and show the assault path is closed quite than simply displaying a patch was utilized. Visa calls this Imply Time to Adapt, and the white paper tracks it alongside three dimensions. Stock freshness measures how present and full the group’s view is of code, configuration, and runtime deployment. Exploitable paths per launch counts what number of end-to-end assault chains stay attainable after every launch, not simply what number of findings have been closed. Validation cycle time tracks how lengthy it takes to produce repeatable, evidence-backed proof {that a} repair works and stays working in manufacturing.
That distinction issues as a result of legacy measures similar to imply time to detect and uncooked CVE closure counts can look higher on paper whereas precise publicity retains rising beneath them. A corporation can shut a whole bunch of findings a month and nonetheless depart viable exploit chains open if no person examined whether or not the patches truly break the assault. MTTA forces groups to measure the final result that issues, and the white paper leans on CISA Identified Exploited Vulnerabilities information to make the prioritization case, noting that fewer than 1% of CVEs are ever actively exploited. Visa’s SSDLC coverage now assumes each exploitable path can be exercised in manufacturing and requires it to be remediated before code is promoted.
Provide chain threat accelerates beneath AI
The dialog moved previous Visa’s personal perimeter when Taneja turned to suppliers. A well-defended enterprise stays uncovered via weak distributors and weak open-source elements, the white paper warns, so Visa is making AI-specific safety posture a non-negotiable dimension of provider due diligence, with expectations for steady vulnerability validation, dwelling software program payments of supplies, and MTTA baselines throughout its know-how stack.
Visa has additionally joined Undertaking Lightwell, the $5 billion IBM and Red Hat initiative to harden extensively used open-source elements via AI-driven validation and coordinated patching, alongside monetary establishments together with Financial institution of America, JPMorganChase, Goldman Sachs, and Mastercard. The dedication extends the similar logic upstream, as a result of the MTTA clock does not pause at any single firm’s perimeter.
When brokers begin shopping for issues
Securing agentic commerce is Visa’s subsequent drawback. Taneja described a future the place AI brokers transact on behalf of shoppers and enterprises, and stated Visa is constructing the belief framework, id layer, and agent readiness scoring that retailers will want before brokers can safely full transactions. Behind that work sits the Visa Fee Threats Lab, a simulation setting the place actual fraud situations get replayed towards the authorization guidelines, thresholds, and configurations Visa truly runs, to floor AI-enabled failure modes as focused hardening suggestions.
The id problem is not theoretical. VentureBeat’s Pulse analysis discovered that 69% of enterprises already run credential sharing someplace of their agent deployments, and corporations with shared credentials report safety incidents or near-misses at a 63.5% charge, towards 40.9% the place each agent has its personal scoped id. Visa’s white paper addresses that hole immediately, itemizing “AI brokers are identities” amongst its 12 non-negotiable practices and requiring scoped permissions, least privilege enforcement, full audit trails, and inclusion in IAM governance for each agent that calls an API, reads information, or modifies a system.
Three priorities for defenders
Visa is organizing its defensive technique round three priorities, Taneja stated. Shift safety left till exploitable flaws are designed out before they attain manufacturing, and exchange high-risk, under-supported elements before they flip into materials publicity. The third is the heaviest raise at Visa’s scale, refactoring defenses to run autonomously beneath human governance so detection, validation, and response hold tempo as menace quantity grows and the fashions behind assaults enhance.
None of it requires a fee community’s price range to begin. The harness sits on GitHub with 595 stars and 97 forks as of July 20, MTTA wants a dashboard quite than a procurement cycle, and the white paper’s 12 non-negotiable practices map onto structure opinions safety groups already run. Visa’s personal conclusion reads like a deadline. The opening to get forward of machine-speed attackers is nonetheless there, the paper argues, and it’ll not keep open.
Disclaimer: This article is sourced from external platforms. OverBeta has not independently verified the information. Readers are advised to verify details before relying on them.
