Download the PHP package promptphp/intercept-tool-approval-guard without Composer
On this page you can find all versions of the php package promptphp/intercept-tool-approval-guard. It is possible to download/install these versions without Composer. Possible dependencies are resolved automatically.
Download promptphp/intercept-tool-approval-guard
More information about promptphp/intercept-tool-approval-guard
Files in promptphp/intercept-tool-approval-guard
Package intercept-tool-approval-guard
Short Description Tool approval guard middleware for Laravel AI agents.
License MIT
Homepage https://intercept.promptphp.com
Informations about the package intercept-tool-approval-guard
Introduction
ToolApprovalGuard is a Laravel AI SDK agent middleware that inspects the tool calls an agent proposes while pausing for human approval, before those calls are ever surfaced for review.
It can block, log, or fully delegate handling to a custom callback.
[!Important] This middleware is part of the Intercept middleware collection. It inspects the tool calls the model proposed. It cannot inspect the tool results, retrieved documents, or conversation history that may have influenced them.
Why this matters
Intercept sees the prompt on its way to the provider. It does not see what a tool returns, what a retrieval step pulled in, or what the model then decided to do with it.
When an agent is manipulated by content Intercept never saw, the damage shows up as a tool call:
ToolApprovalGuard inspects those proposed calls. It is the first point downstream of that blind spot where the middleware pipeline can act.
By default it is the card number that flags this call, not the address. A mail tool is expected to carry an email address, so the defaults cover only values that are essentially never a legitimate argument: card numbers, API keys and bearer tokens. Contact data and locators can be added, but the destination itself is better controlled with allowed_tools / denied_tools and domain allowlisting inside the tool.
Quick start
Installation
You may publish the config or not, the middleware works out of the box.
Usage
Return the ToolApprovalGuard middleware on an agent's middleware method.
[!Important] To add middleware to an agent, implement the
HasMiddlewareinterface and define a middleware method that returns an array of middleware classes.
The middleware only acts when a run pauses for approval. Agents without approval-gated tools are unaffected.
Restricting which tools may be proposed
An empty allowedTools permits every tool. A non-empty list permits only those named.
Observing before enforcing
For the complete guide, see the full documentation below.
Documentation
Full documentation can be found at https://intercept.promptphp.com/ or the docs directory on GitHub.
Contributing
Thank you for considering contributing to Intercept by PromptPHP. The contribution guide can be found in CONTRIBUTING.md.
Code of Conduct
We follow the Laravel Code of Conduct. We expect you to abide by these guidelines as well.
Security Vulnerabilities
If you discover a security vulnerability within Intercept by PromptPHP, please email Victor Ukam at [email protected]. All security vulnerabilities will be addressed promptly.
License
Intercept by PromptPHP is open-sourced software licensed under the MIT license.
Support
This library is created by Victor Ukam with contributions from the Open Source Community. If you've found this package useful, please consider sponsoring this project. It will go a long way to help with maintenance.
All versions of intercept-tool-approval-guard with dependencies
laravel/ai Version ^1.0
promptphp/intercept-injection-guard Version *
promptphp/intercept-pii-redactor Version *
promptphp/intercept-support Version *