deliverability
Skill1 starsChanged 47 days agoMITView on GitHub
Use when auditing email deliverability, checking SPF, DKIM, DMARC, BIMI, MTA-STS or TLS-RPT records, running a pre-send QA gate on an email before it ships, scoring email copy for spam-filter risk, diagnosing why mail lands in spam from message headers, classifying bounces for suppression, checking an IP or domain against DNS blocklists (Spamhaus, Barracuda, SpamCop), planning IP or domain warmup, or producing a deliverability report. Wraps the `deliv` CLI from email-deliverability-suite.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- vercel-labs/marketing-team-eve-templatetheir own version, first seen Oct 23, 2026
- 808enzo/chappietheir own version, first seen Sep 17, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- vercel-labs/marketing-team-eve-template★ 459same name
- 808enzo/chappie★ 3same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 98
Before answering, ask for the ESP's DKIM selector and rerun with --selector.
- Distinctive: line 131
Treat exit code 1 (REVIEW) as a failure, or allow it, according to the team's policy.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Email deliverability with deliv
deliv is a tested CLI with eight subcommands. Each one prints a readable report, and with --json it returns
machine-readable output you can parse. Run the commands. Do not recreate their logic by hand.
Setup
deliv --version || pipx install git+https://github.com/andrewluxem/email-deliverability-suitePick the command
| The user wants to... | Run |
| Check a domain's email authentication DNS | deliv dns-audit <domain> --json |
| Know if copy will trip spam filters | deliv spam-scan <file> --json |
| Gate an email before sending (go / no-go) | deliv copy-check <file> --json |
| Know why a delivered message failed auth or landed in spam | deliv header-analyze <file.eml> --json |
| Decide whether to suppress a bounced address | deliv bounce-classify "<bounce text>" --json |
| Check whether an IP or domain is blocklisted | deliv blacklist <ip-or-domain> --json |
| Ramp volume on a new IP or domain | deliv warmup-plan --esp gmail --start-volume 50 --target-volume 50000 --days 30 --json |
| Hand someone a single shareable report | deliv report --input results.json --format html --output report.html |
Useful flags:
dns-audit --selector <name>probes the ESP's DKIM selector as well as the common defaults. If DKIM comes back WARN with "no key found", ask the user for their selector (their ESP's DNS setup page lists it) and rerun.dns-auditandblacklistaccept--timeout <seconds>(default 10) and--nameserver <ip>.blacklist --no-verifyskips the test-entry check. Keep it on unless you are rate-limited.copy-checkandspam-scanread.html,.txt, or.eml, or-for stdin. Prefer an.emltest send when one is available, because only.emlinput lets copy-check verify the List-Unsubscribe headers.bounce-classify --batchreads one bounce per line from stdin.
Interpret the results
Grades (dns-audit, spam-scan, header-analyze, blacklist): A 90+, B 80-89, C 70-79, D 60-69, F under 60. A failed
critical check always caps the grade at F: a missing SPF or DMARC record, SPF +all, or a broken SPF record.
Each finding is pass, warn, fail, or info. info is never scored. Every warn and fail carries a
remediation string. Quote it rather than paraphrasing it.
copy-check verdicts and exit codes:
PASS(0): safe to send.REVIEW(1): a human should look at the listed reasons first (e.g. no physical address detected).FAIL(2): block the send. Fix every reason listed before retrying.- Exit code 3 means the input file could not be read. It is not a verdict.
blacklist statuses: listed (fail, with a delisting URL), clean, error (timeout, rate limit, or refused
resolver), unverified (the list did not confirm its own test entry, so its "clean" cannot be trusted), and
skipped (a retired list). Never tell the user an IP is clean on Spamhaus when the status is unverified.
Tell them to rerun with --nameserver pointed at their own resolver.
bounce-classify: classification is hard, soft, block, or unknown, and action is suppress, retry,
or investigate. block means the problem is on the sending side (reputation, authentication, content), so never
suppress the recipient for a block. Check confidence: low means only a bare SMTP code matched.
warmup-plan output is planning guidance, not a guarantee of inbox placement. Say so when you present it.
Report back
- Lead with the grade or verdict and the one-line
summary. - List every
fail, then everywarn, each with itsremediation. - For several checks, save each
--jsonoutput and build one report:
deliv dns-audit example.com --json > dns.json
deliv copy-check campaign.html --json > qa.json
deliv report -i dns.json -i qa.json --format html -o deliverability-report.htmlWorked example 1: DNS audit
User: "Why is example.com mail going to spam? Check our DNS."
deliv dns-audit example.com --jsonThe (abridged) result:
{"tool": "dns-audit", "target": "example.com", "score": 76, "grade": "C",
"findings": [
{"check": "SPF lookups", "status": "warn", "message": "SPF uses 9 of 10 DNS lookups. One more sender breaks it.",
"remediation": "Flatten or remove includes before adding another sending service."},
{"check": "DKIM", "status": "warn", "message": "No DKIM key found at the 7 common selectors probed.",
"remediation": "Find your ESP's selector in its DNS setup page and rerun with --selector <name>. ..."},
{"check": "DMARC policy", "status": "warn", "message": "Policy is p=none: you get reports, but spoofed mail is still delivered.",
"remediation": "Review rua reports, fix any legitimate senders that fail, then move to p=quarantine and later p=reject."}]}Before answering, ask for the ESP's DKIM selector and rerun with --selector. If DKIM is actually missing, that is
the most likely cause of the spam placement. Then answer: "Grade C. Three fixes, in order: publish DKIM for
example.com, trim SPF includes (9 of 10 lookups used), and move DMARC from p=none to p=quarantine once reports are
clean."
Worked example 2: pre-send gate
User: "Can we send spring-sale.html tomorrow?"
deliv copy-check spring-sale.html --json; echo "exit=$?"The result is "verdict": "FAIL" with exit=2, and these reasons:
Spam score: Spam-risk grade F (48/100): copy is likely to be filtered.
Link shorteners: 1 shortened link(s) hide their destination: https://bit.ly/3xYz.
Unsubscribe link: No visible unsubscribe link in the body.Answer: "Not yet. The gate fails on three blockers: replace the bit.ly link with a direct link to your domain, add an
unsubscribe link to the footer, and rewrite the top spam phrases." Then list the rewrites array, e.g. "Make money
fast" (line 5): remove it and describe the real offer. After the user edits the file, rerun copy-check and confirm
it returns exit 0 (PASS) or 1 (REVIEW, with the reasons explained) before calling it ready.
In CI
- run: pipx install git+https://github.com/andrewluxem/email-deliverability-suite
- run: deliv copy-check emails/campaign.html # fails the job on FAIL (exit 2)Treat exit code 1 (REVIEW) as a failure, or allow it, according to the team's policy.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

