Mock Recall & Readiness Drill Logger

The problem this kills

If your site holds a GFSI certificate - BRC, SQF, FSSC - or sells to any retailer with a quality program, you have to prove your recall procedure actually works. Not on paper: in practice. That means running periodic mock recalls where you pick a lot, trace every place that product went, reconcile the quantities, and show you accounted for a high enough percentage fast enough.

In most plants this lives in a binder and a panicked spreadsheet. Someone scribbles a start time, chases shipping records, eyeballs the math, and writes "PASS" at the bottom. Then the auditor asks for last year's three drills and the improvement actions you committed to - and nobody can find them, or the numbers don't add up, or there's no proof a manager ever signed off.

The drill itself is fine. The evidence is the weak point. You need timing you can trust, math you didn't fudge, a clear pass/fail against a target you set in advance, and a record a person actually approved - kept somewhere an auditor can see it.

What you'll build

A small, private web app - just for your quality team - that runs a mock recall the right way:

• It picks a random lot from your shipment data so nobody can quietly choose an easy one.
• It starts a timer when the drill begins and stops it when your team reports back, so the trace time is measured, not guessed.
• Your team enters how much product they accounted for versus how much shipped, and the tool computes the percent reconciled and a pass/fail against the target time and target percentage you set before starting.
• If the drill misses target, it captures improvement actions with an owner and a due date.
• Your quality manager reviews and approves the result before it becomes evidence - nothing is finalized until a person signs off.
• It emails a clean summary of the drill and exports the whole record as a CSV for your audit file.

What's inside the Implementation Plan

The plan is a complete, paste-and-go runbook for an AI coding agent (Claude Code). You do not write code - you paste the plan, answer some questions, and approve what it builds.

Crucially, the plan opens by interviewing you about your business. Before it builds anything, the agent asks how your lots and shipments are actually structured, what your traceability target is, who runs drills and who approves them, your typical and peak volumes, and your messy edge cases (split lots, rework, returns, partial shipments). It reflects a short tailored spec back to you, you give a thumbs-up, and only then does it shape the data model and every later step around your operation - not a generic template.

Inside you get: the discovery interview, an account/setup checklist, a plain-language architecture sketch, and the build delivered step by step - each step ending in a ready-to-copy prompt. It closes with how to confirm the tool works and a no-integration fallback that imports a Sheet/CSV and exports clean CSV, so you can build the whole thing today without touching your ERP.

The governance it includes (this is the point)

This is not a toy. The plan bakes in the controls an auditor expects:

• Login so only your team can open the tool.
• Row-level security so you only ever see your own organization's drills - never another site's.
• A complete audit trail - who started the drill, who entered results, who approved, and exactly when.
• A hard human-in-the-loop approval gate - the tool computes the result and drafts the record, but a quality manager must review and approve before it counts as evidence.
• Duplicate guards - each drill has a unique drill ID, so the same exercise can't be logged twice and inflate your records.

Who it's for

Quality managers and food-safety / quality teams who are required to perform mock recalls - for GFSI standards like BRC and SQF, for FSMA, or for customer and retailer audits - and who are tired of reconstructing the evidence after the fact. If you can fill in a spreadsheet, you can build this.

You've got this - paste the first prompt and let the agent interview you.