Limit an AI agent by controlling what its tools can reach—not by asking the model to behave. Give it only task-specific tools, restrict each tool to named resources and operations, isolate code execution, keep long-lived secrets out of its environment, and have an independent policy layer authorize each sensitive action.
Why an agent’s request is not permission
An AI agent can propose a tool call, but the model’s request should not itself authorize that call. A separate execution component should check the actor, tool, target resource, parameters, and any required approval before acting. Unknown or unclassified high-risk actions should be denied by default. OWASP recommends minimum tool access and independent authorization checks for consequential actions in its AI Agent Security Cheat Sheet.
As an Amazon Associate I earn from qualifying purchases.
A prompt such as “do not read secrets” is not a security boundary. If a process can access a file, application, or credential, instructions alone cannot reliably prevent the agent from using it. As OWASP puts it: “Without sandboxing, a compromised agent context has the same privileges as the developer.” (Secure Coding with AI.)
Recommended Free Tools
Start by inventorying what the agent can reach
Make a list of every capability available to the agent, including tools, connected apps, mounted directories, shell commands, APIs, network routes, and credentials. For each item, document its purpose, owner, data sensitivity, and permitted operations. This inventory makes hidden or inherited access visible before you narrow it.
#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.
- Include indirect access, such as a shell that can invoke cloud tools or a mounted home directory that contains SSH keys.
- Record whether each capability is read-only, can modify data, or can send information outside your environment.
- Identify who owns each permission and how it can be reviewed, revoked, or reset.
Scope tools and connected apps to the task
Grant each agent only the tools it needs for its current job. OWASP’s concise guidance is: “Grant agents the minimum tools required for their specific task.” Make scope specific to both the resource and the operation: for example, allow reading a named reports directory while denying writes and excluding secret-bearing paths.
Apply the same principle to app integrations. Authorize access to the particular user, project, or record needed for the request rather than granting an all-purpose app token. Keep read-only agents separate from agents that can edit records or send messages; different trust levels should have different tool sets.
Rank #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
Review MCP integrations before enabling them
Treat Model Context Protocol (MCP) servers as third-party software in the execution path. Maintain an allowlist of approved servers and tools, review their descriptions, pin tool definitions or detect changes, and validate arguments before execution. Do not let an agent discover and connect to arbitrary servers automatically. OWASP discusses these controls in Secure Coding with AI.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Isolate files and code execution
If an agent can run code, assume that code can read, change, or transmit anything its process can access. Run it in a restricted shell, sandbox, virtual machine (VM), or ephemeral cloud workspace. Choose the boundary based on what the task needs to reach; no one isolation option is best for every setup.
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
- Use path and command allowlists, and block sensitive locations such as SSH keys, cloud CLI configuration, environment-secret stores, production credentials, and deployment keys.
- Do not mount broad host directories into the agent’s workspace. Give it only the project files needed for the task.
- Deny outbound network access when it is unnecessary. When access is required, allow only the necessary destinations and observe attempted connections.
- Set limits for CPU, memory, disk, and process count to constrain runaway work.
- Prefer an environment that can be reset or discarded after the task.
A sandbox contains a process; it does not make excessive permissions safe. A broad host mount or powerful credential passed into an isolated runtime can still give the agent broad authority. OWASP’s coding-agent guidance covers runtime, filesystem, and network boundaries.
Keep credentials out of the agent’s context
Do not put a developer’s long-lived credentials, SSH keys, cloud tokens, or organization-wide secrets in the agent’s prompt, retrieved context, tool arguments, or plain-text logs. When an agent needs privileged access, have an execution layer use narrowly scoped authority rather than handing a broad credential to the model.
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.
Prefer task-specific identity and short-lived credentials that expire automatically. For sensitive operations, issue access just in time and for the shortest workable duration, then revoke it when the task ends. OWASP’s Secure Coding with AI and Securing Agentic Applications Guide 1.0 describe ephemeral credentials and just-in-time access as controls.
Require approval for consequential actions
Classify actions by their risk and reversibility. Read-only inspection within an approved scope may need no per-action prompt. Actions that write data, send external messages, change permissions, deploy software, make payments, or delete data in bulk deserve stronger checks or explicit human approval.
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.
- Show the user a preview of the action and its target.
- Bind approval to the actor, tool, target, normalized parameters, timestamp, and expiry. A generic “approved” flag detached from the action is not enough.
- Have the execution service independently recheck policy and approval immediately before acting.
- Use replay protection where relevant, and deny the action if policy lookup, approval validation, or required audit logging fails.
This keeps the approval attached to the specific action the person reviewed, rather than allowing a later or altered tool call to reuse it. OWASP’s agent security guidance recommends independent checks for high-impact actions.
Monitor and test the boundary
Log tool invocations, decisions, and outcomes with enough structured detail to investigate unexpected access. Redact secrets and sensitive content rather than copying them into logs. Alert on unusual calls, repeated denials, unexpected network attempts, and changes to tool configuration.
Keep repeatable tests for prompt override, unauthorized tool use, privilege escalation, data exfiltration, and policy changes. Run them before production and again after material changes to prompts, tools, memory, retrieval, policies, or providers. OWASP recommends adversarial testing and ongoing review in its AI Agent Security Cheat Sheet.
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 glitchesThe OWASP GenAI Security Project’s Agent Control Standard (ACS), dated September 1, 2026, describes runtime middleware hooks for declarative policies and visibility into agents’ access and actions. Treat it as a standards resource, not a plug-and-play product or proof that a particular implementation is secure.
Quick Recap
A practical access-control checklist
- Inventory every tool, app, file mount, command, network route, and credential.
- Allow only the tools and named resources required for the task; separate read from write.
- Review and allowlist MCP servers, detect tool-definition changes, and validate arguments.
- Run code in a restricted, disposable environment with narrow filesystem and network access.
- Keep long-lived secrets out of the runtime; issue scoped, expiring credentials only when needed.
- Make an independent policy service authorize sensitive actions and bind approvals to exact parameters.
- Redact logs, alert on suspicious activity, and rerun abuse-case tests after configuration changes.
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.
