SandboxReceipt

Who it's for

Mac power users, security researchers, and journalists who today cobble together Little Snitch (network only), LuLu, and manual log stream / fs_usage spelunking to check what an app really does

The problem

Time: verifying one app means hours of firewall popups and log archaeology with no definitive answer; Legal/press: no portable signed evidence exists to publish or hand to Apple when an app lies

How to build it

Menu-bar app: pick an installed app, run a 10-minute instrumented session (Endpoint Security + network extension), get a signed PDF/JSON receipt plus optional one-click upload to a searchable public ledger keyed by app+version; watch-mode re-audits automatically on every app update

How it makes money

$39 one-time for unlimited receipts, $59/yr for watch-mode that re-audits every app update and alerts on new egress or file access — Little Snitch's $59 price point proves this exact audience pays for Mac egress visibility, and free tools (LuLu) produce no signed, label-diffed receipt and no update re-check

Why it doesn't exist yet

Incumbents skip it: Objective Development monetizes per-connection firewall rules and has no incentive to maintain a public ledger naming misbehaving vendors, and Apple cannot build it because it becomes Exhibit A against its own review process. Indie gap: Endpoint Security entitlements are now obtainable by solo devs (Objective-See proves the model), and the QBittorrent moment shifted demand from vibes to reproducible proof.

First users

The 1181-point HN thread itself: the macOS forensics and Objective-See-adjacent community who immediately want to reproduce the QBittorrent finding against the other 40 apps on their machines

Build size

1 person x 10-12 weeks: menu-bar app, ESF-based file/network tap, entitlement + privacy-label parser, receipt signer, minimal public ledger site; excludes iOS, excludes real-time firewall blocking, excludes any kernel extension

Biggest risk

Apple denies or revokes the Endpoint Security entitlement for a tool that publishes unflattering evidence about App Store apps, or ships a native label-vs-behavior privacy report in the next macOS

Conditions for a hit (all 3 required)

• Given any installed macOS app, a 15-minute watched session produces a receipt listing every file path read outside the app's container and every remote host contacted, each attributed to a process ID
• Receipt includes a side-by-side diff of declared entitlements and App Store privacy-label categories versus observed behavior, with mismatches explicitly flagged (e.g., label says 'Data Not Collected' but an analytics host was contacted)
• Each receipt is cryptographically signed and published to a public, searchable ledger keyed by app bundle ID + version, so a stranger can verify or compare receipts for the same app version

How it's judged (in 6 months)

GitHub repo with >= 2,000 stars, or Product Hunt daily top 5, or a public ledger hosting >= 200 published app-version receipts(judgment date 2027-03-10)

AI self-confidence 48/100 — self-reported likelihood of meeting the criterion, not a business success rate