iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Stop the agent first, then cut off its access, preserve evidence, determine exactly what changed, and recover through a verified procedure. Do not rely on the agent’s own response to a request such as “undo that” or “stop.” Use an operator-controlled runtime, identity provider, network control, job scheduler, or integration setting. After containment, involve your security or incident-response owner if production systems, sensitive data, credentials, permissions, customer-facing services, or external systems were involved.
This sequence applies whether the agent edited code, changed cloud resources, modified records, sent messages, altered permissions, or triggered another automated workflow. The exact restore command depends on the affected platform; there is no universal rollback command for every agent, repository, database, cloud service, or IDE.
Immediate response: the first 15 minutes
- Pause or isolate execution outside the model. Stop the run, disable the write-capable tool or integration, suspend the queued job, or isolate the workload. Check for parallel workers, scheduled tasks, webhooks, and other agents that could continue the same action.
- Revoke the agent’s ability to authenticate. Disable its token, session, delegated OAuth grant, API key, or service credential. If it used a shared human account, ask the identity owner to review and terminate that account’s active sessions and unnecessary grants.
- Preserve the current state and logs. Record what happened before editing, restoring, or deleting anything. Export relevant audit records to protected storage, restrict access, and redact secrets.
- Open an incident record and assign an owner. Note the start and stop times, affected systems, suspected data sensitivity, and who is authorized to make recovery decisions.
Interruption, rollback capability, least privilege, and independent authorization are core controls in the OWASP AI Agent Security Cheat Sheet. The US Department of Energy’s GEAR guidance likewise emphasizes that authorization must be enforced by the application or infrastructure, not inferred from model output.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware match1. Contain the agent without asking it
An agent can continue through a queued tool call, a retry policy, a background worker, or a second connected agent even after its chat window appears idle. Containment therefore has to operate below the model.
#1 Best Overall
- Powerful Turbo Fan:CAGIWIRU Electric Air Duster have reached speed up to 130000RPM,Can efficient remove dust and debris.With LED light and Three Adjustable Speed to meet with different cleaning tasks.
- With Attached USB C Fasting Charging cable.Charing 3 hours,Enjoy up to 240 minutes of use on the lowest setting, For third speed can use for 25mins.with four charging options to meet with your needs.
- Wide Application: attached 5 different nozzles,making it suitable for a variety of scenes, such as car,pet, pc, keyboards, and other electronic devices. It also serves for office and home clean duster.
- Handy Design & Portable:Cordless air duster electric nozzle is Light weight and small size, design and comfortable to hold and space saving, convenient to carry around. You can put it in your room or inside the car without taking up space.
- Economical &Reusable: Built durable materials with a long-lasting battery,our CAGIWIRU air duster it's a sustainable alternative to disposable air cans.
Runtime and workload controls
- Stop the run from the orchestrator, task queue, container supervisor, CI/CD system, or operating-system process manager.
- Disable write methods while leaving read access available for investigation, if your platform supports separate permissions.
- Block outbound destinations, webhooks, or integrations that could deliver more changes.
- Pause schedules and retries, and identify duplicate runs using the same job, token, or workspace.
- Isolate a host or workload when you cannot confidently separate the agent from other processes.
Document the control used and its timestamp. CISA’s May 1, 2026 guidance on adopting agentic AI services stresses monitoring, threat modeling, layered defenses, and security assessment; these controls are more reliable than a conversational promise from the agent.
2. Revoke credentials and sessions
Containment stops execution you can see. Credential revocation prevents a still-running process, stolen token, or delayed request from authenticating again.
Identify every identity involved
- Agent-specific API keys, service accounts, workload identities, and short-lived tokens.
- Delegated OAuth grants and refresh tokens issued to an integration.
- Browser or IDE sessions, SSH keys, cloud role sessions, and database credentials.
- Shared human accounts used as a shortcut for the agent.
- Credentials passed to downstream tools or other agents.
Prefer an identity dedicated to the agent and limited to the smallest required scope. That lets you disable the agent without disrupting unrelated users. OWASP’s guidance on excessive agency recommends least privilege and user-context identity, while DOE GEAR recommends short-lived, scope-limited credentials and a known revocation process. DOE summarizes the boundary clearly: A prompt such as “never delete files” is not an access control.
See DOE GEAR AI Security and Safety.
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #2
- Screen Cleaner is Your Screen’s New Bestfriend! - EVEO is proud to present a screen cleaner that’s perfectly suited for your TV. Meet your new screen cleaner, computer screen cleaner, laptop cleaner, iPad cleaner, Macbook,tv cleaner or any other electronic device. Our screen cleaner spray and wipe along with the included plush microfiber cloth, easily removes dust, fingerprints & other smudges on the screen’s surface. Clean effectively for a streak-free screen
- Gentle on your Screens - Our screen cleaner spray and wipes is compatible to be used even for the most sensitive LCD, LED, CRT, and OLED screens in proper use. Be at ease knowing that this product is compatible with all the major brands: LG, SONY, Samsung, Sharp, iPad, iPhone, Android screens, Kindle, PC monitors - we’ve got you covered! Can be used as LCD screen cleaner, tv cleaner, smart tv screen cleaner, and more
- Premium Super Soft Plush Microfiber Cloth - Included in this efficient screen cleaning kit. This screen cleaner spray and wipe has our signature plush microfiber cloth specially designed to comply with screens and monitors and leave them streak-free. Has been shown to effectively absorbs stubborn stains, zero streaks, and no fiber shedding. Always use this together with our screen cleaner for the best results
- Easy-to-Use Screen Cleaner Kit + microfiber cleaning cloth - The EVEO Screen Cleaning Kit will easily be your new favorite tool when it comes to taking care of the things you love most, quick and easy solution for a streak-free screen. Simply screw on the sprayer gun on the screen cleaner bottle. Spray the cleaning liquid onto the microfiber cloth, then wipe the screen until desired cleanliness & shine is achieved.
- Quality Screen Care Kit Worth Purchasing - Enjoy a convenient Screen Cleaner anytime, anywhere you need it. Your satisfaction is our top priority. We stand behind the quality of our products and that’s why, you shouldn’t be worrying at all. We manufacture the Best Screen Cleaner. and can be used for laptop cleaning kit, laptop screen cleaner, tv screen cleaner for smart tv. MacBook screen cleaner, monitor cleaner or screen cleaner spray for electronic devices with microfiber cleaning cloth
After revocation
Confirm that the credential is actually invalid by checking the identity provider, token-introspection result, or service audit log. Rotate secrets that may have been exposed, but preserve the old credential’s metadata and revocation time in the incident record. Do not paste live secrets into tickets, chat, screenshots, or copied log files.
3. Preserve evidence before changing the state
Recovery can destroy the clues needed to understand the incident. Capture evidence while it is still available and keep an untouched copy under access control.
Record the action timeline
- When the agent started, when the unauthorized action occurred, and when containment happened, with timezone.
- Agent, model, workflow, workspace, and run identifiers, if available.
- Tool and integration names, versions, parameters, destinations, and responses.
- Prompts or source inputs that influenced the action, including attached files or retrieved context.
- Human approvals, denials, policy decisions, and any changed approval state.
- Errors, retries, partial results, and the final state reported by each tool.
- Resources touched, identities used, permissions granted, and external recipients.
DOE GEAR identifies workflow versions, provenance, actions, approvals, timestamps, outputs, and human review as useful evidence. Preserve the original log format where possible, apply retention rules, restrict who can read it, and redact credentials or unnecessary personal data. Hashing or write-once storage can help show that an exported record was not altered, but use the controls already approved for your organization.
Rank #3
- 【Large size】：Measuring 800 x 300 mm/ 31. 5×12 inches, which is wider and taller than others. Large mouse pad can perfectly fit your keyboard and mouse of any size, and there is also ample space for peripherals such as phones, tablets, etc.
- 【LED backlight mouse pad】：Blade Hawks RGB mouse pad provides 7 static modes and 7 dynamic modes will give you the options you are looking for. Power-off memory function remembers the last lighting mode you selected. The super glow fiber Chroma will give you a colorful gaming setup that you want.
- 【Anti-slip rubber base】：Made from Natural rubber, this gaming mousepad will firmly grip your desktop, allowing you to enjoy the ultimate precision in the games you are most passionate about. It also comes with a texture that enhances this ability even further.
- 【Smooth and waterproof surface】: Micro-textured cloth surface, for extremely precise mouse control and a smooth movement. When liquid splashes on the gaming mouse pad, it forms water droplets and slides down. Will not delay your work or gameplay.
- 【Plug and play】: This mouse and keyboard pad is powered by USB, plug and play, no driver required. One button control light modes. Press one time to change the lighting mode, press twice to change the brightness, press and hold about 3 seconds turn ON/OFF.
4. Establish what changed and the blast radius
Do not roll back merely because a change looks suspicious. First compare the current state with a trusted baseline, version history, audit trail, backup, or last approved deployment. Mark facts as confirmed, suspected, or unknown.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Check more than the original resource
- Data: records created, modified, deleted, exported, or disclosed; include caches, snapshots, and backups.
- Access: new users, roles, policies, tokens, keys, firewall rules, or ownership changes.
- External actions: email, chat, tickets, public posts, purchases, deployments, or API calls to another organization.
- Infrastructure: compute, storage, DNS, network, configuration, billing, and production release changes.
- Downstream automation: webhooks, scheduled jobs, plug-ins, and other agents triggered by the first action.
OWASP lists data exfiltration, privilege escalation, high-impact changes, and cascading failures among agent risks. Microsoft’s guidance on reducing autonomous agentic AI risk recommends post-execution logging and isolation to limit blast radius; review Microsoft’s agentic-risk controls for the relevant architecture.
Decide whether the event is still active
Search authentication, tool, network, and application logs for the revoked identity and related destinations. Look for retries or activity from a different credential. If you cannot establish that writes have stopped, keep the system isolated and escalate rather than beginning restoration.
Rank #4
- 【Softer and More Comfortable】Vaydeer wrist rest has unique diamond pattern, which is the combination of softness and aesthetics. The materials of wrist rest are improved into higher quality memory foam and covered with silky smooth lycra. The computer wrist rest makes you as comfortable and cushiony as like rest your wrists on clouds.
- 【Ergonomic Wrist Saver】The wrist rests for keyboard and mouse comes with a 17.32×3.15×0.83 inch keyboard wrist pad and a 5.94×3.15×0.83 inch mouse wrist support. Based on ergonomic design, the unique concave shape is the perfect fit for your wrist joints. The wrist rest pad fits most computer keyboards and laptops, improve hand and wrist posture, release your wrist and arm stress.
- 【Non-Slip Rubber Bottom】Featuring an anti-skid silicone base on the bottom, this wrist keyboard support stays firmly in place on your desk, preventing the padding from sliding around, ensuring stable and consistent wrist support during extended computer sessions.
- 【Better Experience & Pain Relief】Our keyboard arm rest is beneficial to alleviate the soreness caused by direct contact and friction between your arm and a hard desk surface, reducing the risk of wrist fatigue or carpal tunnel. The soft texture of memory foam can evenly distribute the pressure around your wrists and provide good support with just enough give.
- 【Helpful in Multiple Scenarios】Whether you're working, studying, writing, typing, gaming, this keyboard and mouse rest combo is an essential accessory to add comfort and support to your hands and wrists. It’s also a great gift for men, women, family, friend, coworker, gamer, teacher, etc.
5. Recover through a platform-specific, verified process
Once evidence is preserved and scope is understood, choose the least destructive action that returns the system to an approved state. A rollback can overwrite legitimate work made after the unauthorized change, break dependencies, or erase evidence.
Choose the recovery type
- Revert a single versioned change when the affected object has a trusted history and dependencies are understood.
- Restore from a known-good backup or snapshot when many objects changed or the current state cannot be trusted.
- Apply a compensating change when a direct revert would remove valid later work; document the new state and why it is safe.
- Rotate and rebuild when credentials, permissions, or infrastructure configuration may have been exposed rather than merely changed.
Before execution, record the target, operator, approval, expected result, dependency checks, and a way to stop. Use the affected service’s current official recovery documentation and your incident-response owner for exact commands. The general guidance supports rollback capability but does not define one recovery procedure for every cloud provider, database, repository, IDE, or agent product.
Verify after recovery
- Confirm the restored configuration, data integrity, permissions, and service health against the approved baseline.
- Check dependent services, queues, integrations, and customer-facing functions.
- Review logs for new writes or authentication attempts during and after restoration.
- Preserve the pre-recovery state and the recovery operator’s commands or change record.
6. Escalate and communicate at the right threshold
Contact security or incident response when the agent touched production, sensitive or regulated data, credentials, access controls, customer-facing services, an external system, or when the scope is uncertain. Include your timeline, affected identities, evidence location, containment status, and known or suspected data exposure.
Best Value
- 【No Drill Mounted】Cable management tray can be assembled in as little as 3 minutes or less-simply. Side clips easily clampe the edge of desk, pass through the holes on the sides, and secure excess cables with baskets. Desk cable management supports inward or outward installation to meet your needs in different scenarios, and provides you with great convenience when collecting and organizing wires.
- 【Rubber Pad Protects Desk】Wire organizer under desk built-in soft rubber pads on the clamp protect your desk from scratches, so you will not damage your furniture or office.
- 【Sturdy and Beautiful】Cord management under desk is made of high-quality carbon steel and holds about 15 lbs. The wire basket has 2 holes on each side of the tray to enable your cables to pass freely. Cable baskets are suitable for desk thickness use from 0.4" to 2.4".
- 【Space Saving & Keep Organized】Desk cable management tray holds all your power cords, outlet strips, USB hubs and computer transformers hidden from your feet, and the data cables will not enter the vacuum cleaner to keep your desk clean and tidy. Desk wire organizer is also suitable for use in the kitchen and outdoors.
- 【Reduce Safety Risks】Cable tray under desk keeps all plugs away from your kids, no more worry about tripping. Using 2 holes wire tray saves extra space, helps you easily cabling and wire protection. The mesh design is not easy to accumulate dust, but avoids safety accidents caused by messy wires.
Legal, contractual, or regulatory notifications depend on jurisdiction, sector, contracts, and the facts established during investigation. Do not treat general AI guidance as a breach-notification rule. Let the organization’s legal, privacy, compliance, and communications owners make that decision.
7. Re-enable the agent only with explicit controls
Do not restore the previous broad permissions simply because the immediate change was reversed. Reintroduce capability in stages.
- Assign an accountable human owner who can approve, stop, and review the workflow.
- Start read-only. Confirm that prompts, retrieval sources, and tool routing behave as expected without write access.
- Remove unnecessary tools and destinations. Separate development, staging, and production identities and environments.
- Enforce authorization in deterministic code or infrastructure. A model-generated statement of intent is not an approval.
- Bind approval to the exact action: actor, tool, target, parameters, time window, and expiry. Require fresh approval when any of those change.
- Keep an operator-controlled stop and rollback path. Test it before granting additional write scope.
- Review action logs continuously and set alerts for privilege changes, unusual destinations, high-volume writes, and failed authorization.
These controls reflect recommendations from the OWASP Gen AI Security Project’s Excessive Agency guidance, the OWASP agent security cheat sheet, and Microsoft’s risk-management guidance.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
How to compare recovery and control options
When several recovery paths are available, evaluate them against the same questions rather than choosing the fastest-looking button.
Quick Recap
| Decision axis | What to ask | Safer characteristic |
|---|---|---|
| Reversibility and evidence | Can you undo the action, and will the method preserve the original state and logs? | Snapshot or versioned recovery with an untouched evidence copy |
| Blast radius | How many resources, dependencies, users, and external systems will the operation affect? | Targeted change tested against dependency history |
| Access revocation | Can you disable the agent without interrupting unrelated workflows? | Dedicated, scope-limited identity and short-lived credentials |
| Approval precision | Does approval name the exact actor, tool, target, parameters, time, and expiry? | Fresh, action-bound approval enforced outside the model |
| Auditability and ownership | Who authorized, executed, and verified the action, and where is the record? | Named human owner, protected logs, and a documented verification step |
Common mistakes that make the incident worse
- Asking the agent to undo itself: the same faulty permissions, context, or tool route may still apply.
- Deleting logs or immediately restoring a backup: you may destroy evidence or overwrite legitimate work.
- Rotating only the visible API key: refresh tokens, delegated grants, sessions, and shared accounts may remain active.
- Assuming the first changed resource is the only one: inspect downstream automations, external calls, and permission changes.
- Re-enabling broad access after a manual fix: the underlying authorization gap remains.
- Using a generic rollback command: platform semantics differ, and a destructive restore can create a second incident.
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.
