Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchPC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Make the restriction part of the agent’s runtime, not just its instructions. Give it only the file tools, directories, credentials, and network access its task needs; isolate code execution; and require approval for sensitive actions. A prompt saying “don’t open that file” is not a filesystem permission.
Why prompts alone cannot protect files
An agent can use tools, run code, or pass model-generated arguments to other services. If its process or a connected tool can read a file, the model’s promise not to read it does not remove that access. External content—such as a web page, email, or document—can also contain instructions intended to manipulate an agent. OWASP identifies prompt injection, tool abuse, privilege escalation, and data exfiltration among agent risks. Treat external content and tool arguments as untrusted, and enforce permissions in the runtime or downstream service.
As an Amazon Associate I earn from qualifying purchases.
See OWASP’s AI Agent Security Cheat Sheet and Secure Coding with AI guidance.
Secure the agent in six steps
1. Inventory everything it can reach
List the agent’s file-reading and writing tools, shell commands, MCP servers, mounted folders, credentials, and network destinations. Include extensions and integrations: a tool server may have broader machine access than the file tool shown in the agent interface. Review tool descriptions and configuration changes as part of the trust boundary; do not assume the agent touches only files named in the task. OWASP’s Secure Coding with AI guidance discusses reviewing these access paths.
#1 Best Overall
- POWERFUL SECURITY KEY: The Security Key C NFC is the essential physical passkey for protecting your digital life from phishing attacks. It ensures only you can access your accounts.
- WORKS WITH 1000+ ACCOUNTS: Compatible with Google, Microsoft, and Apple. A single Security Key C NFC secures 100 of your favorite accounts, including email, password managers, and more.
- FAST & CONVENIENT LOGIN: Plug in your Security Key C NFC via USB-C and tap it, or tap it against your phone (NFC) to authenticate. No batteries, no internet connection, and no extra fees required.
- TRUSTED PASSKEY TECHNOLOGY: Uses the latest passkey standards (FIDO2/WebAuthn & FIDO U2F) but does not support One-Time Passwords. For complex needs, check out the YubiKey 5 Series.
- BUILT TO LAST: Made from tough, waterproof, and crush-resistant materials. Manufactured in Sweden and programmed in the USA with the highest security standards.
2. Remove unnecessary tools and scope the rest
Disable capabilities the task does not require. Prefer narrow operations over a broad shell or generic file API—for example, a read-only tool that can access one project subdirectory rather than arbitrary files. Enforce an allowlist of permitted paths and operations at the tool boundary, and bind authorization to the initiating user or session.
Validate normalized paths and arguments before execution; reject traversal attempts and anything outside the allowed scope. Deny-by-default authorization is stronger than relying on a list of sensitive filenames. Blocking patterns for environment files, keys, certificates, or files with “secret” in their names can add defense in depth, but cannot replace an allowlist and operating-system permissions. OWASP’s agent security guidance illustrates scoped, read-only access alongside blocked patterns.
3. Isolate shell commands and generated code
If the agent can run commands or generated code, run it as a restricted operating-system user in an isolated shell, development container, virtual machine, or ephemeral cloud workspace. Do not mount a home directory, unrelated repositories, SSH keys, cloud CLI configuration, or production secrets into that environment. Use read-only filesystems and resource limits where practical.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsRank #2
- POWERFUL SECURITY KEY: The YubiKey 5C NFC is the most versatile physical passkey, protecting your digital life from phishing attacks. It ensures only you can access your accounts
- WORKS WITH 1000+ ACCOUNTS: Compatible with popular accounts like Google, Microsoft, and Apple. A single YubiKey 5C NFC secures 100+ of your favorite accounts, including email, password managers, and more
- FAST & CONVENIENT LOGIN: Plug in your YubiKey 5C NFC via USB and tap it, or tap it against your phone (NFC), to authenticate. No batteries, no internet connection, and no extra fees required
- MOST SECURE PASSKEY: Supports FIDO2/WebAuthn, FIDO U2F, Yubico OTP, OATH-TOTP/HOTP, Smart card (PIV), and OpenPGP. That means it’s versatile, working almost anywhere you need it
- PRIMARY & SPARE KEYS: Just like having a spare house key, we recommend buying two YubiKeys - one for daily use and one as a spare. That way you’ll never get locked out of your accounts
Restrict outbound network traffic to approved destinations when unrestricted internet access is unnecessary. Check where each component actually runs: a sandbox does not protect files exposed through a separate file tool or an MCP server outside the sandbox. See OWASP’s guidance and OpenAI’s API sandbox security documentation.
4. Keep credentials out of the agent’s environment
Agent-generated code can read credentials available to its process. Avoid putting long-lived developer credentials in prompts, source files, environment variables, or logs. Where possible, keep application keys outside the execution environment and use a trusted server or proxy to supply narrowly scoped credentials only for approved hosts and operations. Prefer task-scoped or short-lived credentials; rotate or revoke exposed credentials if compromise is suspected.
A secret manager does not protect a secret after it has been injected into an environment that agent-generated code can read. Credential separation and brokering are covered in OWASP’s Secure Coding with AI guidance and OpenAI’s sandbox security documentation.
Rank #3
- POWERFUL SECURITY KEY: The YubiKey 5 NFC is the most versatile physical passkey, protecting your digital life from phishing attacks. It ensures only you can access your accounts
- WORKS WITH 1000+ ACCOUNTS: Compatible with popular accounts like Google, Microsoft, and Apple. A single YubiKey 5 NFC secures 100+ of your favorite accounts, including email, password managers, and more
- FAST & CONVENIENT LOGIN: Plug in your YubiKey 5 NFC via USB and tap it, or tap it against your phone (NFC), to authenticate. No batteries, no internet connection, and no extra fees required
- MOST SECURE PASSKEY: Supports FIDO2/WebAuthn, FIDO U2F, Yubico OTP, OATH-TOTP/HOTP, Smart card (PIV), and OpenPGP. That means it’s versatile, working almost anywhere you need it
- PRIMARY & SPARE KEYS: Just like having a spare house key, we recommend buying two YubiKeys - one for daily use and one as a spare. That way you’ll never get locked out of your accounts
5. Enforce authorization outside the model
Check permissions at the tool gateway or downstream service, not only in a system prompt or the model’s interpretation of what the user wants. Give the agent identity only the resources and actions it needs, preserve the requesting user’s authorization context, and require human approval for sensitive or irreversible operations.
Recommended Free Tools
For example, an email assistant asked to summarize messages generally needs read access; it should not also have send or delete capability unless the task requires it. OWASP describes the risk of excessive agency; AWS Prescriptive Guidance for generative AI agents also addresses agent permissions and security controls.
6. Test the actual boundary with hostile inputs
Use a harmless canary file or blocked directory to confirm that access is denied. Test ordinary requests and adversarial instructions embedded in documents, issues, or web content. Exercise the real runtime and every access path, including MCP integrations; a reassuring model response is not proof that a tool was blocked.
Rank #4
- POWERFUL SECURITY KEY: The Security Key NFC is the essential physical passkey for protecting your digital life from phishing attacks. It ensures only you can access your accounts.
- WORKS WITH 1000+ ACCOUNTS: Compatible with Google, Microsoft, and Apple. A single Security Key NFC secures 100 of your favorite accounts, including email, password managers, and more.
- FAST & CONVENIENT LOGIN: Plug in your Security Key NFC via USB-A and tap it, or tap it against your phone (NFC) to authenticate. No batteries, no internet connection, and no extra fees required.
- TRUSTED PASSKEY TECHNOLOGY: Uses the latest passkey standards (FIDO2/WebAuthn & FIDO U2F) but does not support One-Time Passwords. For complex needs, check out the YubiKey 5 Series.
- BUILT TO LAST: Made from tough, waterproof, and crush-resistant materials. Manufactured in Sweden and programmed in the USA with the highest security standards.
Audit denied attempts, tool calls, and file changes. Repeat structured tests before production and after material changes to tools, prompts, memory, retrieval, policies, or model providers. OWASP’s AI Agent Security Cheat Sheet and VS Code security documentation cover testing and security considerations.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What VS Code’s controls do—and do not do
VS Code documents that built-in agent tools can read and write within the current workspace folder by default. Additional folders can be granted read-only access through a setting, and the Tools picker can enable or disable individual capabilities. Session permissions can be temporary.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
VS Code separately documents OS-level agent terminal sandboxing as Preview on macOS, Linux, and WSL2, and Experimental on Windows. Availability and behavior can change; check the current VS Code security documentation for your platform. Terminal sandboxing is distinct from the agent’s file tools, which use VS Code’s permission system directly, so assess both boundaries rather than assuming one covers the other.
Best Value
- Security Key : Protect your online accounts against unauthorized access by using FIDO2 and U2F authentication with T110. It's the world's most protective security key that works with windows, Mac OS, Linux as well as Chrome, Firefox, Edge and many other major browsers.
- Certified with the new FIDO2 standard, T110 provides the benefit of fast login and strong protection against phishing, account takeover as well as many other online attactks.
- Works with : Bank of America, Github, Google, Microsoft, DUO, Twitter, Facebook, Dropbox, Apple, ebay, BINANCE, mor and more.
- Fits USB-A port : Insert the T110 security key into the USB-A port of each service and log in conveniently with one touch
- For the driver download and user guide, please visit TrustKey Solutions Home support page.
How to judge whether a setup is actually restrictive
“Sandboxed” or “safe” is not enough detail to assess a configuration. Check the enforcement point, scope, execution location, credential exposure, network reach, operational friction, and whether access attempts can be logged and retested.
- Enforcement boundary: Is access blocked by operating-system permissions, a sandbox, or a tool gateway—or only discouraged by model instructions?
- Scope: Can rules limit paths, operations, users, and sessions?
- Execution location: Do the shell, file tool, MCP server, and remote connector all run inside the same restricted environment?
- Credentials and network: Are secrets kept outside the agent environment, and can the agent contact only necessary destinations?
- Operations: Can denied attempts be audited and the controls retested after configuration changes? Do approval steps protect high-impact actions without blocking routine work?
No single control guarantees prevention across every agent architecture. The practical objective is to limit what each component can reach and verify that the restriction holds through the tools the agent actually uses.
Quick Recap
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.
