Think-Act Security & Controls Access · Approval · Review

Control what the AI agent can read and change.

Looking up a price, filling in a form, and submitting it are different actions. Tell Think-Act where to stop, and check the tool permissions that determine whether an action proceeds, asks for approval, or is blocked.

Windows 10 and 11 now. Mac release planned.

Agent settings include controls for how tools may act.

Controlled task workflow

Capability Approval Review
Agent settings include controls for how tools may act.
Data flow & processing

Where does the information in a task go?

The app runs on your desktop. A cloud model or connected service may receive the text, images, or files needed for a request.

Review Integrations
01

Context

Limit the inputs

Give the agent the relevant folder or documents rather than everything on your desktop.

02

Think-Act

Specify what may change

Ask for a list of proposed renames first if you want to check them before editing files.

03

Service

Check the destination

A local app can still send text to a model provider or a connector. Check both before supplying private data.

04

Review

Verify the result

Open the resulting file. If the agent changed a website or another app, check the change there as well.

The key point: running Think-Act locally does not mean every service involved in a workflow also runs locally.

Security model

Allow, ask, or block

Allowed actions can run without a prompt. Ask waits for you; Block prevents the action.

The goal is deliberate access, not friction everywhere

Low-impact work can stay simple. Actions with meaningful side effects deserve more scrutiny while a workflow is being understood and tested.

Capability

What can this workflow actually use?

Keep the run tied to supported tools, Skills, and connectors that are actually configured and available to the task.

Permission

Should this action proceed automatically, pause, or stay blocked?

For supported actions, Think-Act can follow configured ALLOW, ASK, or BLOCK policy. Match the policy to the action and its consequences.

Review

What can you inspect before accepting the result?

Review the final output together with relevant supported activity, tool use, observations, confirmations, or reported changes. Treat this as review evidence, not a complete audit trail.

External trust

What sits outside Think-Act's own controls?

Third-party MCP packages, model endpoints, runtimes, accounts, and connected services bring their own behavior and data-handling rules. Review them independently before relying on the workflow.

Agent settings include controls for how tools may act.

Permission is action-specific

Do not assume every connector exposes one identical permission surface. Use the controls supported by the capability you are configuring.

Agent settings include controls for how tools may act.
Tool permissions

Enable the tools the job needs

A read-only comparison doesn’t need permission to change a business record. Start with the smaller scope.

Allow

Proceed under the configured policy

Useful when the scope and consequence are understood and an approval pause is not required.

Ask

Pause for a decision

A practical starting point for write, send, publish, delete, or other consequential actions while testing a workflow.

Block

Keep the action unavailable

Use when the action does not belong in the task or its consequence is not acceptable.

Keep the capability set small

Disable or leave out connectors the task does not require where appropriate.

Keep consequential actions deliberate

Increase autonomy only after you understand the action, its expected side effects, and how you will verify success.

Skills do not grant missing access

A Skill preserves reusable guidance. It does not create a connector, bypass missing capability, or override the workflow's permission boundary.

Local MCP risk

Treat local MCP packages as third-party code

Check the publisher and what the package can access. Running locally is not a guarantee of trust.

Third-party MCP packages are not automatically trusted

Third-party MCP packages are not controlled or audited by PixLab unless explicitly stated. Think-Act's workflow permission model does not replace the operating-system permissions or security properties of the package itself.

Review Local MCP Setup
Approval & review

Check the target before you approve

Read the filename, destination, or requested action. Pausing a task doesn’t reverse a change that has already happened.

Completion is not proof

Compare important claims, values, extracted details, file changes, or external effects with the source and the outcome you expected.

  1. Request

    A supported action becomes part of the run

    The workflow can only rely on supported capabilities that are actually available to the task.

  2. Review / decide

    Pause when the configured policy requires approval

    Inspect the requested action and consequence, then approve or deny it deliberately instead of treating confirmation as a routine click-through.

  3. Inspect

    Review available activity and tool evidence after the run

    Use relevant supported activity, observations, confirmations, or reported changes to understand how the workflow reached its result. This is review evidence, not full chain-of-thought visibility.

  4. Recover / reuse

    Recover where supported, then improve the method before repeating it

    Use supported checkpoints or recovery controls where they apply. External actions may not always be reversible, so do not treat recovery as universal undo.

    Turn a workflow into a reusable Skill only after the prerequisites, access, expected result, and review checks are understood.

Extend a workflow deliberately

A Skill cannot override an approval

Saved instructions can standardize a task, but they don’t change tool access or waive your permission rules.

Keyboard reachable

Important controls, dialogs, approvals, and navigation should remain usable without requiring precise pointer input.

Clear labels

Actions should explain what they do so users can distinguish approval, cancellation, review, retry, and other choices.

Readable contrast

Text, controls, focus indicators, and important state changes should remain distinguishable across light and dark interfaces.

Warnings with context

A warning should identify the issue and help explain why an action is blocked, risky, unavailable, or requires confirmation.

More than color

Important states should communicate meaning in more than one way

Approved, blocked, unavailable, warning, and error states should combine meaningful text, recognizable icons, clear labels, and sufficient visual contrast rather than relying on blue, green, amber, or red alone.

Approved Review Blocked
Current integration model

Verify requirements before a rollout

Check the settings available in your build. Ask PixLab about organization-wide controls rather than assuming they come with a plan.

Reusable Skills Configured connectors Local MCP Trust & review
Try Think-Act with Local MCP

Planned / roadmap

Roadmap ideas stay separate from current product behavior

Broader managed integration concepts may include cloud connector discovery, organization-level policies or provisioning, and remote-agent integration ideas.

Cloud connector discovery
Organization policies & provisioning
Remote-agent integration concepts
Review Current Product FAQ
Security FAQ

Agent security questions

Find out when an approval appears and how local and external tools handle data.

Are third-party MCP tools automatically safe?

No. MCP packages can run code with the permissions of the process. Use only packages from sources you trust.

Does desktop mean everything stays local?

No. Local files can still be sent to a configured model or connector when needed for a task. Check those destinations before sharing private data.

What determines whether an action needs approval?

The existing security policy considers the requested tool and action. It can allow it, ask you to confirm it, or block it.

Can a Skill bypass permissions or add external access?

No. Reusable instructions do not grant permissions, bypass blocked actions, or install missing tools.

What happens when a required connector is unavailable?

Inspect the failed step. Enable the right trusted connector or choose a result that doesn't need it.

Can every Think-Act workflow run offline?

No. Web research and hosted models require a connection. Local processing depends on your configured provider and tools.

How do I review what happened during a run?

Open the task's activity log and the resulting files. If the job changed another app, check the result there as well.

Which security capabilities are current and which are planned?

The app has action approvals, protected paths, and tool activity records. Don't assume a centralized enterprise policy or audit dashboard is included.

Bounded workflows · Windows & macOS

See what a read-only run looks like

Ask the agent to check the free space on your drives. Inspect the tool activity before giving it a job that changes files.

Scoped capabilities Deliberate approval Review before relying