Nova Mcp Laravel Package

Product Decisions This Supports

• Build vs. buy for AI/admin interoperability: v1.2.0 adds a prompts resolver, meaning the package can now expose predefined MCP prompt templates alongside tools/resources. This turns "should we build our own MCP bridge for Nova?" into a clearer buy decision—prompt scaffolding is now handled by the package rather than bespoke code.
• Roadmap sequencing for AI-assisted operations: Prompts support unlocks guided, repeatable AI workflows (e.g., "summarize this customer's recent orders," "draft a moderation response") over Nova data. This is typically the difference between a demo and a shippable internal AI feature, so it can justify pulling AI-admin work forward a quarter.
• Feature scoping for internal tooling: PMs can scope an "AI copilot for back-office staff" MVP using Nova resources + MCP prompts, rather than committing to a separate RAG/agent platform.
• Differentiation and support burden: Assess whether prompt templates should be owned by product (curated, versioned prompts) or left to platform engineering; the resolver gives a natural extension point for product-managed prompt content.
• Risk/security review trigger: Exposing prompts and Nova data to MCP clients expands the attack surface; this release is a prompt to schedule a permissions and data-exposure review before rollout.

When to Consider This Package

Consider adopting when:

• You run Laravel Nova as the admin surface and want AI clients (Claude Desktop, IDE agents, internal assistants) to interact with admin data through a standard protocol.
• You need prompt templates plus tool access, not just raw CRUD tools—v1.2.0's promptsresolver makes curated, reusable AI workflows feasible.
• Your team is PHP/Laravel-fluent and prefers extending an existing package over standing up a separate Node/Python MCP service.
• You want a thin, incrementally adoptable integration: tools first, prompts and richer workflows later.

Look elsewhere when:

• You are not on Nova (or are migrating off it)—the value is tightly coupled to Nova's resource model.
• You need fine-grained, audited multi-tenant authorization beyond what Nova policies provide; you may need a custom MCP layer or additional gateway controls.
• Your AI use cases require heavy retrieval, embeddings, or agent orchestration—this package is a bridge, not an AI platform.
• You require stable, long-term APIs and cannot absorb the churn of a young, fast-moving package (v1.2.0 in early 2026; commit history suggests active iteration).
• Regulatory/data-residency constraints forbid exposing admin data to external MCP clients without a dedicated security architecture.

Adoption criteria to clear before committing: Nova version compatibility, MCP client support matrix, auth model for MCP connections, prompt template ownership/governance, and a documented upgrade path for minor releases.

How to Pitch It (Stakeholders)

For executives: This package lets our existing Laravel Nova admin securely expose data and guided AI workflows to AI assistants via the Model Context Protocol—a standard, not a bespoke integration. The new v1.2.0 prompts feature means we can ship curated, product-owned AI workflows (not just data lookups) on top of an admin we already run, reducing build cost and time-to-value for internal AI tooling. Recommendation: pilot with a narrow, low-risk workflow and a security review before broader rollout.

For engineering: Adopt opscale-co/nova-mcp as the MCP bridge for Nova. v1.2.0 adds a prompts resolver for loading MCP prompt templates, giving us a clean extension point for versioned, product-managed prompts. Evaluate Nova version constraints, MCP client auth, and policy coverage early; plan to own prompt content and a permissions review. Expect active upstream iteration—pin versions and track releases.