we wrote the playbooks. once. properly.

every dtc team rebuilds the same 24 marketing playbooks. ad copy frameworks, klaviyo flows, pdp audits, retention math, attribution setups. we shipped them, productionized them, and packaged them as skills.

track record

  1. 2018-2020
    growth lead, $4M supplement DTC
    scaled paid spend $30k→$180k/mo at 2.4x ROAS across meta + google.
  2. 2020-2022
    head of lifecycle, $12M apparel DTC
    rebuilt klaviyo flows: welcome 1.2%→3.8% cvr, abandon-cart 4%→9%, post-purchase ltv +18%.
  3. 2022-2024
    freelance growth, 14 brands
    47 funnel audits shipped, average 22% ltv lift in 90 days across cohort.
  4. 2024-now
    building skillor
    24 marketing skills shipped while running 3 client retainers. dogfooded every skill before release.

mission

reduce the time between 'we should test that' and 'we shipped it' to one prompt. marketing infrastructure should be portable and ownable — not locked behind another saas seat that bills monthly forever.

what is a skill

a skill is a single markdown file (SKILL.md) with a frontmatter contract and a process body. drop it into claude code, claude desktop, or any agent harness that reads markdown skills. claude reads the contract, follows the process, returns the artifact. no api. no installer. no runtime. you own the file.
  • frontmatter declares: when to invoke, required inputs, brand context fields.
  • body declares: process, decision rules, output format, edge cases.
  • everything else: yours.

why now

ai agents only do useful marketing work when given opinionated, well-scoped instructions. generic prompts produce generic playbooks. the leverage is in the structure, not the model. skillor is that structure, in 24 pieces, for $99.

skeptical? read one.

READ A SKILL