Killed By Claude Report
← HomeCurrent verdict
Assessment
Anthropic clearly has real momentum in AI-assisted cybersecurity, especially around:
That is bad news for weakly differentiated security software.
We do not know what Wabbi actually sells.
A parked domain and a two-word description are not enough to conclude direct overlap. Anthropic threatens parts of security workflows, not all security companies equally. If Wabbi is actually a governance, compliance, orchestration, or enterprise process layer, Claude is more of a feature than a full substitute.
Right now this looks like adjacent-to-meaningful exposure, not a confirmed wipeout.
If Wabbi's core value is "help security teams analyze code and find/fix vulnerabilities," then the score should be much higher.
If its value is "system of record, policy control, workflow enforcement, or enterprise integration," it still has some room.
Biggest historical hit
The clearest threat is "Claude Code Security" from Feb 20, 2026.
That announcement moves Claude from being a general model with security potential into a named security product direction: scanning codebases for vulnerabilities and proposing targeted patches for human review.
If Wabbi lives anywhere near AppSec, code scanning, or remediation workflow, this is the moment the threat became concrete rather than theoretical.
What still protects them
If Wabbi survives, it will not be because "security software" is a good moat. It isn't.
Real protection would have to come from one or more of these:
If none of that exists, then Claude becomes the smarter front-end and Wabbi becomes replaceable middleware.
If those things do exist, Claude is more likely an engine Wabbi can use than a complete substitute.
Signals
Why this is in the blast radius
This is the most direct overlap signal in the pack.
Anthropic explicitly says Claude can scan codebases for vulnerabilities and suggest targeted software patches for human review. If Wabbi's security software touches AppSec, code scanning, remediation, or developer security workflows, that is squarely in its lane.
If Wabbi is instead focused on compliance, GRC, or unrelated security infrastructure, the overlap drops sharply.
This is strong proof that Claude is not just talking a big game in security.
Anthropic claims Claude found 22 Firefox vulnerabilities in two weeks, with 14 rated high severity. That directly threatens vendors whose value proposition is finding software flaws faster or better.
It does not prove overlap with every kind of security software, but it materially raises risk for any company selling vulnerability discovery or AppSec assistance.
Opus 4.6 explicitly highlights enhanced cybersecurity abilities and defensive use in finding and patching vulnerabilities in open-source software.
That means the security capability is part of the core model stack, not just a one-off demo. For a startup in generic security software, this increases substitution risk because the underlying model keeps getting stronger at software security tasks.
Opus 4.7 improves advanced software engineering and long-running, rigorous task execution. That matters for security because real remediation work is not just finding bugs—it is tracing impact, editing code, verifying fixes, and handling messy repos.
This is not a security-specific launch, so the overlap is indirect. But it expands Claude's ability to do the surrounding engineering work that security products often wrap.
Computer use helps Claude operate across live apps and multi-step workflows. That could let Anthropic span ticketing, IDEs, dashboards, and remediation systems without bespoke product UX.
But this is still an enabling capability, not a dedicated security product by itself. It raises platform risk more than it proves direct replacement of Wabbi.