All packsTransactional Email Deliverability CoreSolve
email suppression list complaint overwritten by bounce
A complaint is not a bounce that arrived later
A naive suppression upsert lets a later hard-bounce webhook overwrite a spam-complaint row with a weaker reason -- silently un-suppressing a reputation signal.
This is one of the things Transactional Email Deliverability Core already handles. Send transactional email that actually lands, and stop emailing people who bounce or complain.
Is this you?
A spam complaint and a hard bounce are both reasons to suppress an address, so it's tempting to store both in the same suppressions table with a plain upsert -- whichever event arrives last wins. If a hard-bounce webhook for the same address arrives after (or is reprocessed after) a complaint, a naive upsert silently downgrades the row's reason from complaint to hard_bounce, erasing a signal that's supposed to be legally and reputationally permanent.
Why this one is easy to get wrong
An upsert (ON CONFLICT DO UPDATE) is the standard, idiomatic way to keep a suppressions table current, and "the most recent event wins" is a reasonable default rule for almost every other field in a system like this. Complaints being a one-way door -- never downgradeable, regardless of what arrives afterward -- is a domain rule that has to be stated explicitly; it doesn't fall out of normal upsert semantics.
What you get instead
Suppression decision 2 makes the upsert conditional: ON CONFLICT ... DO UPDATE only when the existing reason is not complaint. Decision 3 backs this up on the read side too -- unsuppress() refuses to remove a complaint-reason row at all, returning {removed: false} and leaving it in place, because a complaint is treated as a reputation signal to preserve, not a typo to fix.
Source: ARCHITECTURE.md 'Suppression semantics', decision 2 — checkable in the pack you receive
How you actually use this
You don’t install a library or wire up an SDK. Your own coding agent builds the code in your project, and you keep it — no runtime dependency on us.
Step 1
Download and unzip
You get a folder: the docs that tell an agent what to build, a starting skeleton, and the test suite that decides when it's done.
Step 2
Open it in Claude Code or Cursor
Point your coding agent at the folder. Nothing to install, no account with us, no API key.
Step 3
Paste one prompt
The pack contains the exact prompt. Paste it as your first message and leave it alone — it works through the build itself, choosing a cheaper or stronger model per task.
Step 4
Run ./verify.sh
One command. It prints a pass or fail for every check. Green means the build is done — the same script we ran to produce the receipt on this page.
Typical build: about 13 minutes of your agent working, mostly unattended. Then you integrate the working module into your app the way you would any code you’d written yourself.
Why you can believe this
3 of 3 runs passedWe ran this pack from an empty folder 3 times and published exactly what happened — every check, the model, the token cost, the wall time. Not a testimonial, and not our opinion: the same verify.sh you run yourself. Read the full receipt →