Spintax makes every send of a variant read slightly differently so a thousand copies of one email don't pattern-match as a thousand copies of one email. This skill defines what's safe to spin (wording with no unique meaning) and what's never spun (numbers, names, merge tags, the case study, the offer's unique mechanism), plus when to move from light spin at launch to deep spin once a variant crosses roughly 2,000 cumulative sends.
Pipeline Playbooks · Open Source, MIT Licensed
Rules for spinning finished cold email copy so a thousand sends of one variant don't look identical, what to spin, what to protect, and when to deepen it.
Every skill folder is self-contained. Run this from your project root to pull just this one into .claude/skills/spintax-depth/:
npx degit joelwylie7/cold-email-playbooks/.claude/skills/spintax-depth .claude/skills/spintax-depth
Want the whole system, not just one skill?

Two base cold email frameworks, Offer First and Built For You, plus poke and hook opener layers and the hard bans that keep copy out of spam.

Joel's 4-step cold sequence: a first touch, a same-thread nudge at +3 days, a fresh thread with a new offer angle at +30 days, then a final nudge.

A numbered gate that checks copy, settings, leads, senders, schedule and sequence structure before any campaign launches or resumes, PASS or FAIL, one line of evidence each.