laravel-evidence-risk-review banner

Laravel Evidence Risk Review

Evidence-boundary labeling and risk-sweep guardrails for Laravel AI products, RAG systems, review pipelines, and MCP tools.

Documentation site: https://doc.laravel-evidence-risk-review.padosoft.com (Cloudflare Pages custom domain; pending DNS/deploy verification before public launch)

Quickstart

Install the package, publish config, and run your first deterministic review.

Open ->

Architecture

Understand why every public surface delegates to the same ReviewEngine.

Open ->

Reference

Use the PHP, CLI, HTTP, MCP, and config reference pages.

Open ->

What This Package Does

It answers a narrow but important question: not just whether a citation exists, but whether the cited evidence is strong enough for this claim, in this domain, for this population or boundary condition.

flowchart LR A[Artifact] --> B[Evidence tier labeling] B --> C[Risk sweep] C --> D[Verdict reduction] D --> E[ReviewResult] C --> F[(Review log)] G[Domain profile] --> C H[Budget meter] --> C

Core Guarantees

  • Standalone Laravel package: no host application coupling.
  • HTTP, MCP, LLM, and persistence are default-OFF.
  • No hard LLM SDK dependency.
  • One core engine behind PHP, Artisan, HTTP, and MCP surfaces.
  • Append-only review log when persistence is enabled.

Mathematical Sketch

A simplified risk score can be read as a normalized aggregate over per-claim severity and evidence gap:

R=1∣C∣∑c∈Cseverity(c)⋅gap(c) R = \frac{1}{|C|} \sum_{c \in C} severity(c) \cdot gap(c)

See Risk Score for the deeper model.

next Quickstart