Back to vendors
A

AWS Continuum

Also known as: AWS Security Agent, Continuum for penetration testing, Continuum for code vulnerabilities

Visit site
Entry price$200 per 1 million lines at the first code tier plus inference, or $50 per testing task hourFull pricing detail

AWS Continuum is an Amazon Web Services product for application security teams. It combines code vulnerability management with penetration testing for applications the customer controls. The code service uses a graph of repositories and AWS resources to prioritize findings and delivers proposed patches into developer workflows. The testing service keeps the scope of an application in an Agent Space and can recheck a specific finding after a patch goes live.

AWS Continuum is an Amazon Web Services security service that covers code vulnerability management and application penetration testing. The code service works with existing findings, scans connected source, and ranks issues by the role of the affected systems in the business. It brings validated issues to developers with evidence and proposed fixes. Penetration testing, previously called AWS Security Agent, covers applications under the customer's control and keeps the application's test scope and security requirements in an Agent Space.

The code service groups repositories, context and scan policy in an environment. It uses a graph of AWS resources, network reachability and identity relationships to assess findings against the deployment, and confirms exploitability in an isolated sandbox with reproduction steps. GitHub repositories or code stored in Amazon S3 can provide the source. The service can run scans every day, every two days, once a week or on demand.

The Issues view brings together severity, contributing findings, validation evidence and patch details. Scans can finish before issue processing does, and the console distinguishes scan status from the status of each proposed fix. Continuum can deliver a proposed change to a connected repository for review. People decide what merges.

Penetration testing starts from the target URLs, credentials and context a team supplies, including design documents, user stories and infrastructure as code. Its agents run attack scenarios that chain several steps, and every action lands in a live log, with the HTTP requests and responses behind each finding kept as evidence.

Tests run once, on a daily, weekly or monthly schedule, or from a CI/CD pipeline that gates deployments in GitHub Actions, GitLab CI/CD, Bitbucket Pipelines or Azure DevOps. With write access switched on, the agent opens fix pull or merge requests across seven source control platforms, files Jira issues for findings and publishes reports to Confluence. Teams can cap a run at a maximum number of task hours and recheck individual findings after deploying a fix.

The code service operates in Northern Virginia and Oregon. It keeps stored data in the region where a request starts, with inference able to run in another US region. AWS Identity and Access Management governs access to Continuum resources, and the service sits in AWS's SOC 1, SOC 2 and SOC 3 scope under the AWS Security Agent name. AWS manages the model portfolio for code security work and selects models by task.

That service bills for the amount of code protected and the inference used: 1 million protected lines with 100 vulnerabilities processed in a month, at one inference unit each, comes to $1,200. Penetration testing bills $50 per task hour, measured by the second, and is included in the AWS Free Tier plan for new AWS customers, who also receive up to 400 task hours each month for two months. Parallel tasks accrue time separately, so the meter can advance faster than a wall clock.

The code service fits application security programs that need to prioritize many repository findings in the context of an AWS deployment. Its testing counterpart can also cover authorized applications in other clouds, hybrid setups or on premises. Luminate Data uses the code service's attack path visualization to discuss specific risks with product and business stakeholders.

Infor uses it to reach multiple frontier security models through one service, and IBM and Tata Consultancy Services are also named customers. HENNGE reported a reduction of more than 90% in its typical testing duration using the testing product. The two capabilities have distinct billing units, so a team that runs both can separate recurring code protection from individual testing runs.

Vendor details

Canonical URL

https://aws.amazon.com/continuum/

Category

Security / SOC agent

Subcategory

Application security validation and remediation

Company status

first party product

Use cases & customers

Primary use cases

Prioritize application vulnerabilities using environment contextReview proposed fixes in development workflowsRecheck a remediated finding after release

Target customers

Application security teamsCloud security engineersDevSecOps and vulnerability management teams

Deployment options

Code protection in US East Northern Virginia and US West OregonAuthorized application testing across AWS, other clouds and on premises

Integrations

Penetration testing connects to GitHub, GitHub Enterprise Server, GitLab in the cloud or self managed, Azure DevOps, and Bitbucket Cloud and Data Center for code context and fix pull requests, to Jira Cloud for finding issues and to Confluence Cloud for context and reports, and runs from CI/CD pipelines in GitHub Actions, GitLab CI/CD, Bitbucket Pipelines and Azure DevOps. The code vulnerability service draws on GitHub repositories, Amazon S3 code sources and AWS Config collectors. AWS Identity and Access Management, the AWS CLI and AWS SDKs govern service access.

In practice

An application security lead groups an application's repositories into a Continuum environment and reviews the evidence behind critical issues in the issue queue.

A developer receives a proposed dependency update as a pull request and checks its effect before approving the change.

A release team deploys a fix and uses the testing workspace to recheck the recorded finding and preserve the result in its history.

Agentic Index coverage score

11.5 / 14 capabilities · 82%

Integrations & Tool Calling Full

Penetration testing connects to GitHub, GitHub Enterprise Server, GitLab and GitLab Self-Managed, Azure DevOps, and Bitbucket Cloud and Data Center. Once write access is switched on for a repository, the agent opens pull or merge requests with fixes and posts review comments. A Jira Cloud connection lets it create an issue for a finding and add comments as the finding is rechecked, and a Confluence space can receive published reports. The code vulnerability service delivers its patches to GitHub, takes source from Amazon S3 and gathers resource data through AWS Config collectors.

Sourcedocs.aws.amazon.com/securityagent/latest/userguide/about-integrations.htmlread 2026-11-20

Workflow Orchestration Full

A code vulnerability environment coordinates scanning, validation, business risk assessment and patch generation. Findings from source analysis are linked to an issue and can continue through remediation processing after the scan itself finishes. The penetration testing service coordinates specialized agents in parallel and brings their results together under the application scope defined for that run.

Sourceaws.amazon.com/continuum/faqsread 2026-11-20

Knowledge Grounding & RAG Full

The security graph ties protected code to the resources and dependencies in the customer's AWS environment. It brings together resource topology, identity relationships and network reachability so that the service can connect a code finding to the deployed application and its business context. AWS maintains the graph as part of ongoing protection and uses it to assess issues as the environment changes.

Sourceaws.amazon.com/continuum/pricingread 2026-11-20

Human Oversight & Guardrails Partial

Operating policies control whether a code environment generates patches. Agent Spaces give administrators control of application scope and security requirements, and test budgets can stop a run while preserving its findings. For code fixes, Continuum prepares a change and a person reviews and merges it in the repository. Approval for the merge lives in that repository workflow.

Sourceaws.amazon.com/blogs/aws/aws-continuum-for-code-vulnerabilities-is-now-generally-availableread 2026-11-20

Security, Identity & Governance Full

AWS Identity and Access Management controls who can sign in and which Continuum resources each person can use, and federated users can come in through IAM Identity Center. Users see only the Agent Spaces they are assigned to, and test credentials stay in AWS Secrets Manager or a Lambda function. The service appears in AWS's SOC 1, SOC 2 and SOC 3 scope under its former name, AWS Security Agent, linked to the Continuum product page. Customer data is encrypted at rest with AWS KMS and is not used for model training.

Sourceaws.amazon.com/compliance/services-in-scope/socread 2026-11-20

Observability & Auditability Full

A penetration test logs every action the agent takes as it runs, sorted by action type and risk type, with a testing summary and detailed logs of all activity. Each finding carries a trace of the HTTP requests, responses and exploitation attempts behind it, and each task reports its own duration and task hours. Administrators can send run logs to Amazon CloudWatch, and AWS CloudTrail records every API call with the identity that made it. On the code side, the Issues view keeps the contributing findings, evidence and remediation summary for each issue.

Sourcedocs.aws.amazon.com/securityagent/latest/userguide/review-penetration-findings.htmlread 2026-11-20

Memory & State Persistence Partial

Each Continuum environment keeps its scans, findings and issues as work moves from discovery to remediation. A confirmed finding becomes an issue that can hold evidence and proposed patches. The testing service also keeps the history of later checks linked to the original finding in the application workspace. Those records preserve work across runs, with no fixed retention term for the finding history.

Sourcedocs.aws.amazon.com/continuum/latest/apireference/welcome.htmlread 2026-11-20

Deployment & Data Residency Full

The code service is hosted by AWS in Northern Virginia and Oregon, with configuration tied to the region where the customer creates the environment. Stored data remains in the originating region. AWS can move inference between US regions to use available capacity. The separately configured testing service can cover an authorized application hosted in AWS, another cloud, a hybrid environment or on premises.

Sourcedocs.aws.amazon.com/continuum/latest/userguide/data-protection.htmlread 2026-11-20

Prebuilt Agents / Templates / Packs Full

Continuum offers code vulnerability management and application penetration testing as two ready to use capabilities. The code service arrives with an operating sequence that takes findings through risk review and proposed remediation. The testing service supplies its own specialized agents and uses the customer's Agent Space for application scope and requirements. Customers configure each capability around a workload in their AWS account.

Sourceaws.amazon.com/continuumread 2026-11-20

Triggers & Channel Coverage Full

Scan frequency belongs to each code environment. A new environment starts an initial scan, and subsequent scans can run daily, every two days or every seven days. Manual mode keeps the initial run and leaves later work to a person queuing it in the console. An environment carries its configured frequency without someone having to open the Issues view or start each recurring run.

Sourcedocs.aws.amazon.com/continuum/latest/userguide/cv-triggering-scan.htmlread 2026-11-20

Model Flexibility & Routing Partial

The code service uses multiple frontier security models and selects among them by the needs of each task. AWS manages model access and blends models behind a common inference unit. The account pays for the security work rather than separate model subscriptions. The selection stays within the managed service, without a model menu or credentials supplied by the customer.

Sourceaws.amazon.com/continuum/pricingread 2026-11-20

APIs / SDKs / MCP Extensibility Full

The AWS Continuum API exposes the code service's environments, integrations, collectors, scans, findings and issues. Teams can manage the code sources an environment protects and retrieve the results through the API rather than relying on the console alone. The service uses AWS signed programmatic access through the AWS CLI and SDKs alongside the access permissions administrators apply to the console.

Sourcedocs.aws.amazon.com/continuum/latest/apireference/welcome.htmlread 2026-11-20

Testing, Debugging & Optimization Full

Finding rechecks let a security team test a specific change after releasing a fix. The testing service can recheck selected findings from a completed run without running the full application test again. It returns an Active or Resolved status and attaches the history to the original finding, giving the team a record of whether the application change held when assessed by the same service.

Sourceaws.amazon.com/about-aws/whats-new/2026/08/aws-security-agentread 2026-11-20

Browser / Computer-use Not documented

Continuum performs application security testing and repository analysis within a defined target scope. The code service delivers outcomes through issue records and patches to connected repositories. The testing service returns findings for the authorized application in its own workspace. The service does not provide a general browser or desktop session for an operator to delegate interface work to the agent.

Sourceaws.amazon.com/continuumread 2026-11-20

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Pricing

$200 per 1 million lines at the first code tier plus inference, or $50 per testing task hour

Protected lines of code, inference units and testing task hours

Trial available

Included quota

The testing trial includes 400 task hours each month. The first three rechecks of each finding are free.

What is public

AWS Continuum charges $10 per inference unit for code security work, plus monthly context rates per 1 million protected lines. Context costs $200 for the first 10 million lines, $180 for the next 90 million, $160 through 1 billion, $130 through 10 billion and $100 above 10 billion. Penetration testing costs $50 per task hour.

Billing mechanics

Context rates apply progressively. The testing meter adds up the active seconds across all tasks in a run, including tasks that overlap.

Variable cost rationale

Inference work varies by finding complexity, and multiple testing tasks can accrue time in parallel.

Additional watchouts

A test can reach its task hour limit before the same number of hours has passed on the clock.

Overage / add-ons

Testing usage beyond the trial allowance is billed at the standard task hour rate.

Sales call required

No, self serve available

Free / trial

Penetration testing is included in the AWS Free Tier plan for new AWS customers and has a two month trial with up to 400 task hours a month

Lowest paid plan

Usage based code protection and application testing

Commercial notes

Both capabilities have no upfront charge or minimum fee. Penetration testing is included in the AWS Free Tier plan for new AWS customers. Code protection has no trial allowance. At 1 million protected lines and 100 vulnerabilities processed in a month, with one inference unit each, code protection comes to $1,200; 10 million lines and 1,000 vulnerabilities come to $12,000.

Missing data

Inference units required per finding vary with the work performed.

Agentic Index verified 2026-11-20

Alternatives to AWS Continuum

Other products that do the work AWS Continuum does, and what each one is.

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.