Type Guard Laravel Package

Product Decisions This Supports

• Build vs. buy (tooling, not features): Type Guard is not a user-facing capability — it is a developer-ergonomics dependency that makes PHP's type system legible to static analyzers (PHPStan/Larastan, Psalm). The real decision it informs is whether to standardize a "narrow once at the boundary" convention in your Laravel codebase instead of writing bespoke is_string()/assert() ladders in every controller, job, and listener.
• Quality investment on high-cost bug classes: Type coercion and untyped payload handling are disproportionately expensive in billing, entitlements, webhooks, and multi-tenant scoping — places where a string "0" or a nullable field silently changes behavior. Adopting narrowing helpers is a cheap lever to shrink that class of defect.
• Static analysis ambition: If the roadmap includes raising PHPStan to level 6+ or making Larastan a CI gate, this package supports that path by giving analyzers explicit narrowing evidence rather than forcing @var docblock casts and inline ignores.
• Third-party and asynchronous boundaries: Concrete use cases worth scoping: Stripe/webhook payload parsing, queue job payloads, JSON columns in Eloquent, config and env reads, and request data after validation. These are the seams where "trusted type" assumptions leak.
• Buy signal for a thin internal wrapper: The pragmatic PM decision is not "add this package everywhere," but "add it behind a small internal Types/Guard helper" so the team gets the ergonomics without coupling application code to a 198-star, single-org library.

When to Consider This Package

• Your team already runs PHPStan/Larastan (or Psalm) in CI and has hit the level where untyped mixed values generate noise.
• You have a meaningful volume of boundary code — webhooks, queue consumers, API clients, config access — where narrowing happens repeatedly.
• You want a minimal, MIT-licensed, single-purpose dependency with a tiny surface area rather than a full validation framework (this is not a replacement for Laravel's validator or data_get/DTO casting).
• You value low lock-in: a library this small can be vendored, forked, or reimplemented in an afternoon, which materially lowers adoption risk.
• Look elsewhere when: the codebase has no static analysis discipline (the payoff is largely in analyzer signal, not runtime behavior); the team is small enough that native PHP 8 is_* checks plus assert() suffice; you already standardize on DTOs/spatie/laravel-data or Symfony Serializer for typed hydration; or you need strong vendor guarantees, active CVE response, and a support contract.
• Diligence flags: 198 stars and a single maintaining organization mean bus-factor risk. Require a thin internal adapter, pin the version, and confirm compatibility with your PHP/Larastan versions before promoting it beyond a single service.

How to Pitch It (Stakeholders)

• To executives (30 seconds): "A ~zero-cost, MIT-licensed dev utility that helps our type checker catch bugs at the edges of the system — payments, webhooks, background jobs — before they reach customers. No user-facing change, negligible maintenance, easily removed."
• To engineering leadership: "It complements our static analysis rather than replacing anything. It lets us narrow types once at the boundary and have PHPStan honor that narrowing, which cuts defensive re-checking and docblock casts across controllers, jobs, and API clients."
• To security/compliance: "Tiny dependency, single purpose, MIT license, no network or data access. Isolated behind an internal helper so it can be removed without touching product code."
• To the team (adoption ask): "Pilot in one service for two sprints. Success criteria: measurable drop in PHPStan baseline errors, fewer @var/ignore annotations, and no increase in boundary bug reports. If it doesn't move those numbers, we drop it — the exit cost is one afternoon."
• Set expectations honestly: treat this as a 1.5–2.4-tier, nice-to-have quality-of-life dependency, not a platform bet. It earns its place only if static analysis is already a habit.