- What is fissible/verdict used for?
- Verdict is a Laravel package for defining decisions as reusable, testable rules. It answers questions like whether a user can perform an action and replaces scattered conditionals with clear, verifiable verdicts. It is also described as a security boundary for AI-triggered application actions, bounding authority, duplication, rate, and approval.
- How do I install fissible/verdict in Laravel?
- Install the package with Composer using composer require fissible/verdict. Then follow the package README for any service provider registration or configuration publishing steps. The provided excerpt does not list exact installation commands beyond the Composer package name, so check the repository documentation for Laravel-specific setup.
- How do I define a decision rule or verdict?
- Define each decision as a dedicated, reusable rule that answers a single question such as whether a user may perform an action. Keep the rule focused and testable instead of spreading conditionals across controllers or services. The exact class names and method signatures are not included in the provided excerpt, so consult the package source or README for the concrete API.
- What Laravel versions does fissible/verdict support?
- Laravel and PHP version support is determined by the package composer.json constraints. The provided overview does not list specific supported Laravel versions. Check Packagist or the repository composer.json for the exact requirements before installing.
- Can Verdict bound what an AI agent is allowed to do?
- Yes. Verdict is described as a security boundary for AI-triggered application actions. It bounds authority, duplication, rate, and approval rather than trying to control what the agent is attempting to accomplish. This lets you enforce limits even when an agent's goal is otherwise valid.
- How should I test Verdict decision rules?
- Because decisions are defined as reusable rules, you can test them directly with PHPUnit or Laravel's testing helpers. Cover allowed and denied cases for each can-this-user-do-X question and include edge cases around rate, duplication, and approval limits for AI-triggered actions. Keeping tests close to the rule definitions makes changes easier to verify.
- What production concerns does Verdict address for AI-triggered actions?
- Verdict explicitly bounds authority, duplication, rate, and approval for AI-triggered application actions. These controls help prevent an agent from exceeding its authority, repeating the same action, exceeding rate limits, or acting without required approval. The package does not attempt to govern the agent's underlying intent.
- Is Verdict a replacement for Laravel Gates and Policies?
- Verdict is not described as a drop-in replacement for Laravel's Gate and Policy system. It focuses on centralizing decisions as reusable, testable rules and on bounding AI-triggered actions. You can use it alongside Laravel authorization where standard resource policies and explicit verdicts serve different needs.
- Is fissible/verdict actively maintained?
- The repository is listed as unknown in the provided information, so active maintenance cannot be confirmed from this context. Check Packagist, the GitHub repository, and release history for recent commits, tags, and Laravel compatibility updates. For production use, pin a tested version and review the issue tracker.
- What are alternatives to fissible/verdict for Laravel decisions?
- Alternatives include Laravel's built-in Gate and Policy authorization, custom decision or rule classes, and role-based packages such as Spatie permission. Verdict is distinct in framing decisions as reusable verdicts and in adding a security boundary for AI-triggered actions. Choose based on whether you need standard authorization, AI action limits, or both.