The New Stack reports that Continuum became generally available on Thursday, giving developers a service that checks repository vulnerabilities against their AWS environment and proposes code changes.
AWS made Continuum, its superintelligence (SI) code-security service, generally available on Thursday, The New Stack reported on October 24. According to the report, Continuum finds and validates vulnerabilities in code repositories, checks them against the developer’s AWS environment, and opens pull requests—proposed code changes for developers to review.
The release follows benchmark results AWS published on October 5, more than two weeks before launch. AWS reported that Continuum completed 819 of 920 tasks successfully. That 89.0% score measured whether it found and repaired a real vulnerability without breaking existing tests, not whether it repaired a particular bug chosen in advance.
CyberGym-E2E, the benchmark AWS used, gives an agent 90 minutes to find and fix a vulnerability in one of 139 open-source projects. It supplies no bug description, demonstration of the flaw, crash log or existing patch. To succeed, the agent must identify and trigger a flaw, then produce a fix that stops the crash while keeping the project’s existing tests passing.
| Outcome | Continuum | Previous public high |
|---|---|---|
| Repairs a real flaw and keeps existing tests passing | 89.0% | 65.9% |
| Also repairs the benchmark-selected historical vulnerability | 37.8% | 26.2% |
The agent is never told which historical vulnerability the benchmark selected. As The New Stack explains, a repository can contain several vulnerabilities, so repairing a different real flaw helps explain the gap between the scores. The 37.8% figure is a share of all tasks, not a share of successful patches.
Analysis
Every task concerns a memory-safety bug in C or C++: an error in how a program accesses or manages memory. AWS acknowledges that this scope leaves out many other programming languages and vulnerability types. The scores therefore do not measure performance across all the security problems a development team might face.
AWS ran the tests itself. The company says it followed the benchmark’s rules, including restrictions on internet access, and checked that the agent was not simply looking up publicly available fixes. Although the benchmark and testing tools are public, the results had not been independently replicated at the time of The New Stack’s report.
For teams that already know which vulnerability they need repaired, Continuum’s success rate on that specific task also remains unknown. The release makes the reported workflow available: checking repository vulnerabilities against an AWS environment and proposing fixes through pull requests for developers to review.
The W27 keeps the Galaxy Z Fold 8 Ultra’s core hardware, with dark-red or dark-blue finishes, golden accents and 16GB of working memory in both storage versions.
2 min
The satellite operator expects access to more than twice OneWeb’s current bandwidth in return for spending in 2027–2034, with initial IRIS² services expected from 2030.
3 min
The black, gold and green edition follows September’s sold-out preorders. Nintendo’s replenished launch-day stock also quickly sold out.
1 min
