Recommended Free Tools
Contain a compromised AI agent in AWS by stopping the credentials it uses, isolating its host where applicable, preserving evidence, and only then removing the cause and restoring access. Treat the incident as an AWS identity-and-workload compromise, not as a special AWS incident category. Identify the agent’s IAM user, role, or Identity Center principal; determine which resources it can reach; and match containment to the actual credential and runtime path.
This guide assumes the agent is using AWS services directly or is running on an AWS workload such as EC2. The EC2 isolation steps do not automatically apply to Lambda, containers, managed agent platforms, third-party model providers, or credentials issued outside AWS. Those deployments require controls specific to their runtime and provider.
As an Amazon Associate I earn from qualifying purchases.
What an “AI agent compromise” means in AWS
An agent is compromised when an attacker, malicious instruction, stolen token, vulnerable tool, or altered deployment can make the agent act outside its approved purpose. AWS responders should map that behavior to ordinary control points:
- Identity: the IAM role, IAM user, or IAM Identity Center principal whose credentials sign requests.
- Runtime: the EC2 instance, container, Lambda function, managed agent service, or external system running the agent.
- Reachable resources: S3 buckets, databases, queues, secrets, APIs, and other services allowed by policy or network paths.
- Evidence: CloudTrail events, instance data, identity changes, network indicators, and cost or usage anomalies.
AWS’s separate AI Investigative Agent is not the customer agent being contained. It assists AWS-supported Security Incident Response cases by gathering and analyzing evidence. The published guidance describes it as an investigation aid, not as a control that disables or isolates your agent. Human responders remain responsible for evaluating its recommendations.
#1 Best Overall
- [Ultra-Lightweight Ear Buds Designed for Small Ears] Each earbud weighs only 3.7g and features a compact, ergonomic in-ear design made especially for small ears. Secure, low-profile, and comfortable for workouts, all-day wear or overnight sleeping.The ultra-slim profile sits flush in your ear with no hard edges poking or rubbing, so you can even sleep on your side without discomfort.
- [Immersive Stereo Sound with TOZO OrigX Technology] TOZO OrigX tuning delivers clear vocals, balanced mids, and natural stereo sound for music, podcasts, and videos.
- [Long Battery Life for Daily Use] Get up to 7 hours of playtime on a single charge, with up to 32 hours total using the charging case—ideal for workdays, commuting, and extended listening sessions.
- [Bluetooth 5.3 & Stable Connection] Bluetooth 5.3 provides fast pairing, stable wireless performance, and reduced dropouts as you move around home or office.
- [Deep Bass with Clear Vocals] High-performance drivers produce punchy bass while keeping vocals clean and detailed for everyday listening.
“When AWS Security Incident Response identifies an active threat in your environment, containment is the immediate priority: stop the threat actor from causing further damage while preserving evidence for investigation.”
AWS Security Incident Response User Guide
The response sequence at a glance
- Establish scope and coordinate: record the account, region, runtime, principal, resources, and time window.
- Contain the identity: prevent the principal and active sessions from making further AWS API calls.
- Isolate the workload: for EC2, apply a dedicated restrictive security group and assess active connections.
- Preserve evidence: capture relevant memory, disk, logs, and metadata before terminating or rebuilding anything when circumstances allow.
- Protect data and dependent resources: apply narrowly scoped, reversible controls to S3 and other reachable services.
- Investigate and eradicate: trace API activity and permission changes, remove malware or the original access path, and address the root cause.
- Recover deliberately: restore access only after validating identity, host, permissions, monitoring, and business impact.
At each stage, choose the least destructive action that is effective for the current threat. AWS recommends weighing potential impact, evidence and regulatory requirements, service availability, effectiveness, reversibility, and expected duration.
1. Establish scope before changing controls
Create an incident record before making changes, unless an active threat requires immediate emergency containment. Capture enough context for another responder to understand what was changed and why.
Record the identity and deployment
- AWS account ID and affected region or regions.
- Agent name, version, deployment method, and runtime host.
- IAM role, IAM user, or Identity Center user used by the agent.
- Role-session name, access-key identifiers, assumed-role details, and the time suspicious activity began.
- EC2 instance ID, container or task identifiers, Lambda function name, or managed-service resource ID.
- S3 buckets, objects, databases, queues, secrets, and other resources the agent accessed or could access.
- Observed actions, destinations, unusual API calls, new identities, policy changes, and cost or usage anomalies.
AWS incident-response material identifies account IDs, IAM principals, resource identifiers, and a timeframe as useful context when requesting investigation support. If your organization uses AWS Security Incident Response and the case is AWS-supported, open a case with this information. The documented AI investigation operates within that supported-case boundary.
Do not rely on stopping the process alone
Stopping an agent process can prevent new work, but it does not invalidate credentials that were already issued or copied. A running process may also have delegated credentials, long-term access keys, or sessions in another workload. Tie containment to the actual principal and active sessions, then separately address the runtime and any credentials outside AWS.
2. How to stop an AI agent from making AWS API calls
Contain the AWS principal as well as the host. AWS provides the AWSSupport-ContainIAMPrincipal Systems Manager Automation runbook for suspected compromise of IAM users, roles, and Identity Center users. It is designed as a structured, reversible process that retains the identity for investigation. Review the runbook’s current parameters, required permissions, and change plan before execution.
Rank #2
- 【IP68 Waterproof Multifunctinal Smart Watch】Torjalph fitness activity tracker smart watch is IP68 waterproof technology ,you can wear it when swimming,handwashing,it's rain-proof which means that you perform multiple exercise modes outside without worrying about sweat or rain corroding your fitness smart watch,and this watch equipped with a 230 mAh large capacity battery magnetic charging ,and this sports activity tracker watch can be used 5-7 days If fully charged.
- 【Larger 1.83'' Full Screen Watch & Multi-dial Selection】1.83 inch full HD color touch screen smart watch for men / women provides excellent 240*284 resolution image quality ,This Bluetooth smart watch offers a better touch experience and remarkable immersion and ,more than 200 watch faces for you to choose from App"Da Fit ", you can change the dial style according to your mood to wear keeping the world on your wrist
- 【Accurate Activity Tracker Smart Watch 】sports watch activity tracker uses sweat and dust resistant breathable silicone strap, our Fitness Tracker Smart Watch for Android iOS Phones supports 12 sports modes,walking,running,bicycle,swimming,badminton,basketball,football,rope skipping. This bluetooth calls smart watch for men women is lightweight and stylish, bringing more convenient and relaxing sports experience while also providing more protection.
- 【Your 24 Hour Personal Daily Health Monitor】Fitness watch for women men full touch screen IP67 waterproof bluetooth smartwatch for Android iOS Phones with heart rate blood pressure monitor multifunction Andriod smart watch with steps counter sleep monitor sports fitness watch, monitoring blood oxygen saturation allows you to understand your physical condition in real time to track fitness progress ,to know your health status in time.
- 【More Functions Bluetooth Call Smart Watch】This fitness activity smart watch full screen for iPhone and Android phones includes a bunch of new features including:The bluetooth smart watch can hang up the phone, receive and display WhatsAPP /Text /Facebook /Twitter /Messenger reminder.Support sedentary reminder, alarm clock, stopwatch, timer,weather information,music control, remote camera control compatible with Android 5.0 and above or iOS 9.0 and above smartphones.
Choose the control that matches the credential
| Situation | Containment focus | Important consequence |
|---|---|---|
| Agent uses an IAM role | Contain the role and revoke active sessions when necessary. | Revoking sessions affects every user or workload currently using that role. |
| Agent uses an IAM user | Use the IAM-principal containment process and address access keys according to their type. | Long-term access keys are not stopped by the temporary-session timestamp policy described for role sessions. |
| Agent uses an Identity Center user | Use the supported principal-containment workflow and invalidate the applicable sign-in or session path. | Check for other principals or roles the user can assume. |
| Agent has credentials outside AWS | Use the issuing provider’s revocation and rotation mechanism. | The AWS procedures do not provide an exhaustive cross-provider playbook. |
Revoke active role sessions carefully
AWS documents a Revoke active sessions action for roles. It attaches an AWSRevokeOlderSessions policy that denies requests from users who assumed the role before the revocation point. AWS accounts for policy propagation with an approximately 30-second window; sessions assumed more than approximately 30 seconds after the revocation are not affected by that older-session denial.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
This is not a private kill switch for one agent. Every user or workload with a current session from the specified role can be interrupted. Before using it, list legitimate consumers, estimate the service impact, and decide whether a narrower principal or resource control can stop the threat. Keep the role and related logs intact so investigators can reconstruct what happened.
The aws:TokenIssueTime restriction discussed by AWS applies to requests signed with temporary credentials. It does not revoke long-term IAM user credentials. If the agent or attacker has a long-term access key, disable or rotate that key through the applicable IAM-user process, then check for replacement keys, copied secrets, and other roles the principal could assume.
3. Isolate a compromised EC2 environment
For a potentially compromised EC2 instance, AWS GuardDuty and AWS Security Incident Response recommend a dedicated isolation security group. Configure it to allow inbound and outbound traffic only from specifically approved IP addresses, then remove the instance’s other security-group associations. The objective is to keep the instance available for investigation while sharply reducing its communication paths.
Security-group isolation is not a connection kill switch
Changing security groups blocks future traffic; it does not terminate existing tracked connections. An established suspicious flow can therefore continue after the group change. If an active connection must be interrupted, use AWS guidance for enforcing Network ACLs from network indicators of compromise. Record the indicators and the exact rule changes so they can be removed safely later.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesWhen EC2 instructions do not apply
Do not assume an EC2 security group can contain a Lambda function, a container task, a managed agent platform, or a host outside AWS. Those environments have different identity, network, and lifecycle controls. Identify the actual runtime first, then apply its provider-specific isolation method while still containing the AWS principal.
Rank #3
- [Low Noise High Fidelity] New bee guitar cord with 22AWG oxygen-free-copper center conductors provides enhanced signal transfer, the PVC insulation material and the high-density stranded copper double shield provide noise rejection for quiet operation
- [Ideal for Choice] New bee 1/4 straight and right angle metal connectors, ensures a secure connection in any instrument, pedal, or amp jack. L shape great fro on stage guitar stands
- [Wide Compatible] Suitable for mixer, power amplifier, microphone, effector, decoder, equalizer, electoronic organ, electric guitar, electric bass and other electronic equipments. Widely used in the installation of smart home, karaoke speaker, professional stage audio, etc
- [Easy Disassemble] Making incidental repairs quick and easy! Very nice, flexible cable
- [Dont Hesite to Order] 24 hours customer services and professional technology team are standing by
4. Preserve evidence before destructive action
Evidence preservation and immediate risk reduction can conflict. Decide explicitly whether the threat is still active, whether the instance contains unique evidence, and what service outage is acceptable. If the threat cannot be controlled through identity and network measures, safety may require termination even when evidence collection is incomplete.
Use a staged forensic workflow where practical
AWS Prescriptive Guidance describes an automated workflow that acquires memory and disk, then tags an EC2 instance for containment. The pattern stores artifacts in a separate Forensics account and uses a separate Security account for orchestration. This is an architecture option, not a requirement for every incident. Adapt it to your retention, access-control, and regulatory design.
The same guidance cautions that the pattern is not intended to produce artifacts that can be considered electronic evidence submissible in court. If legal proceedings are possible, coordinate collection, custody, and retention with qualified legal and forensic specialists rather than treating an automated workflow as courtroom-ready evidence.
Free tools Windows power users keep installed
One-click scans. No signup required.
Preserve the surrounding record
- Export or protect CloudTrail events for the relevant accounts and time window.
- Record IAM policies, trust relationships, access keys, role-session names, and recent identity changes.
- Capture EC2 instance metadata, tags, snapshots or forensic images, and security-group and Network ACL state.
- Save agent prompts, tool calls, model outputs, application logs, and deployment changes where available.
- Hash or otherwise integrity-protect collected files according to your incident-handling standard.
5. Protect S3 and other reachable resources
Determine whether the agent read, modified, deleted, or exfiltrated data. Review not only the agent’s intended resources but also every resource granted through inherited policies, resource policies, assumed roles, and network paths.
AWS Security Incident Response supports reversible S3 containment by changing access policies to deny external access while leaving objects in place. Scope that action to the affected bucket, prefix, account, or principal. Confirm which legitimate producers and consumers will be interrupted, and avoid deleting objects or logs that may be evidence.
Apply the same principle to other services: prefer a narrowly targeted deny, endpoint restriction, or permission boundary when it stops the unauthorized behavior without taking unrelated workloads offline. If a shared policy or role is involved, document the blast radius before applying a broad deny.
Rank #4
- Digital Family Calendar: BSIMB digital calendar wall touch screen puts an end to your chaotic life. Designed for busy families, the digital calendar keeps you organised and in step with your family. Assign different colours and tasks to each family member to make family scheduling easier and more intuitive.
- Easy Setup and Auto-Sync: Plug in, connect to WLAN and auto-syn your electric calendar in minutes. Digital wall calendar compatible with Google calendar, iCloud calendar, Outlook calendar, Cozi calendar and Yahoo calendar. Add events directly on the device or through the ecalendar app. Electronic calendar stays current whether you're at work or traveling—no manual updates needed.
- All-In-One Home Assistant: Smart calendar builds habits and encourage independence with personalised chores. Display daily meal plan to reduce the daily ‘What's for dinner?’ stress. To-do lists keep track of everything you want. Turning chores into valuable achievements with stars and rewards, making life easier and more enjoyable.
- Manage Anytime Anywhere: BSIMB electronic calendar for family stay connected with the powerful ecalendar app. Access and manage your digital wall calendar anytime, anywhere with free features like stars & rewards, photo screensaver, chore charts, meals plans, to-do lists etc. Electronic planner paid features like Magic Import tailor advanced tools for users with special needs.
- Stay in Touch with Loved Ones: Share photos or videos from your phone to electronic wall calendar with app, instantly update your life moment, and keep in touch with your family, especially those who are far away; you can also surprise your smart calendar digital wall by sending him a wish card; it's also great for keeping tech-illiterate elders in touch with the youngsters.
6. Investigate the cause and eradicate it
Review AWS control-plane activity
Use CloudTrail to reconstruct API calls, source addresses, user agents, session names, and timing. Review IAM for new users, access keys, policy attachments, trust-policy changes, permission escalation, and newly assumed roles. Examine EC2 data and application logs, then check Cost Explorer and service-usage anomalies for unexpected workloads or data transfer.
For AWS-supported Security Incident Response cases, the AI Investigative Agent can query CloudTrail, IAM, EC2, and Cost Explorer and logs its own actions in CloudTrail under AWSServiceRoleForSupport. Treat its output as an aid to triage: AWS states that responders must independently evaluate generated recommendations and maintain human oversight.
Remove the access path, not only the symptom
- Remove unauthorized policy statements, trust relationships, users, keys, schedules, functions, or deployment changes.
- Patch or replace vulnerable agent tools, dependencies, images, and host software.
- Rotate secrets that were available to the agent, including secrets stored in application configuration.
- Reduce permissions to the minimum actions and resources the agent actually needs.
- Check every account and region reachable through the compromised principal or its assumed roles.
For malware on EC2, AWS recommends investigating the instance and removing discovered malware. If unauthorized activity cannot be identified and stopped, AWS recommends terminating and replacing the instance as needed. Coordinate evidence collection and replacement sequencing before taking that irreversible step whenever the threat level permits.
7. Recover only after containment is verified
Restoration is a controlled change, not simply removing the isolation group or re-enabling a key. Confirm that the root cause is addressed and that the replacement identity and workload are clean.
Recovery checklist
- Verify that the compromised principal’s sessions, keys, and delegated credentials are contained or rotated.
- Review the final IAM policy, trust policy, resource policies, and network paths.
- Rebuild an affected EC2 instance from a trusted image when replacement is required.
- Restore S3 or other service access in small, observable steps rather than all at once.
- Enable or validate CloudTrail, GuardDuty, log delivery, alerting, and cost monitoring.
- Run a limited functional test with least-privilege credentials before returning normal traffic.
- Keep heightened monitoring long enough to detect recurrence and document the decision to close the incident.
Should you disable the IAM role or isolate the EC2 instance?
Usually, do both when both control points are available: identity containment stops AWS API calls, while workload isolation limits network communication and preserves the host for investigation. The right order depends on the active threat and service impact.
| Decision axis | Identity containment | Workload or network isolation |
|---|---|---|
| Primary target | IAM user, role, or Identity Center principal | EC2 instance and its network paths |
| Stops AWS API requests | Yes, when the relevant credentials and sessions are covered | Not necessarily; existing credentials may still work through allowed paths |
| Limits network communication | Not by itself | Yes for future traffic; existing tracked connections need separate handling |
| Evidence impact | Retains the identity and its records | Can preserve a live host, but termination destroys volatile evidence unless collected first |
| Blast radius | May affect every consumer of a shared role | May interrupt applications and operators connected to the instance |
| Reversibility | Often reversible, depending on key rotation and policy changes | Security-group changes are reversible; termination is not |
Use the narrowest combination that actually stops the observed behavior. A role used by many production services may require a temporary resource-level deny or replacement role rather than an immediate role-wide session revocation, while an actively communicating instance may require network isolation even after credentials are blocked.
Common failure modes
Only stopping the agent process
Existing role sessions, long-term keys, copied secrets, and other deployments can continue operating. Contain the principal and search for alternate credential paths.
Assuming a new security group ends every connection
Security-group changes affect future traffic, not established tracked connections. Use network-indicator-based Network ACL controls when an active flow must be interrupted.
Revoking a shared role without checking users
All current consumers of that role can be denied. Inventory legitimate sessions and communicate the expected outage before revocation.
Terminating before collecting evidence
Termination can destroy volatile memory and host context. Apply reversible controls first and collect evidence when the threat can be safely contained; terminate when unauthorized activity cannot otherwise be stopped.
Treating generated investigation output as a decision
The AWS AI Investigative Agent can accelerate evidence gathering in supported cases, but responders must validate its recommendations and retain human oversight.
Quick Recap
How to preserve evidence before rebuilding a compromised EC2 instance
- Record the instance ID, account, region, role, security groups, Network ACLs, tags, and exact time of suspected activity.
- Contain the role or other principal so new AWS API calls stop, while considering the effect on legitimate workloads.
- Apply the dedicated isolation security group and remove other associations; address active connections separately if required.
- Acquire memory and disk using an approved workflow when feasible, storing artifacts with restricted access and documented integrity controls.
- Export relevant CloudTrail, IAM, application, agent, and cost records.
- Investigate and remove malware or the exploited access path.
- If activity still cannot be identified and stopped, terminate and replace the instance, then restore service from a trusted build.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
