All articles
The median AI skill ships 2 files. Mine ships 156.
I counted every file in 109 public AI marketing skills. The median ships 2 files, and 91 of the 109 ship no executable code at all. Here is the full tree of mine, with the graders' real PASS and FAIL output.

I read 52 community teardowns. The best champion programs pay nothing.
I expected the best community champion programs to run on budgets. Notion's ambassadors are unpaid. Cash attracts mercenaries; access attracts advocates. Here's that finding, and the harder rule I shipped with it: no champion and no partnership leaves my community skill without a revenue tie or an honest exploratory flag.

I thought every post needed a hook. My best one had none.
I built a skill that enforces hook, structure and CTA on every social post. The rules that turned out to matter are the ones that switch it off: when to drop the machinery entirely, why my buyers read everything and comment on nothing, and why a long-form Article was never just a longer post.

I thought a URL that parsed was a URL that was safe
A URL can pass every parsing rule and still be a phishing domain, a doubly-encoded payload, or a link that silently destroys your ad-cost attribution. Three failures my agents were walking straight past, and the skill I shipped to stop them.

My AI writes the whole cold campaign. I still approve every one.
I wanted an AI that runs cold outreach at volume without firing emails in my name that I never saw. So outbound-engine now builds you a board — it drafts every campaign, then stops at an approval gate you own. Speed of an agent, go-button of a human.

Posting more isn't a social strategy. Running the whole loop is.
Volume isn't a social strategy, and neither is being on six platforms. Here's how one AI picks the channel first, then runs a brand's whole social department one sharpening lap at a time.

58% of my replies come from emails I almost didn't send
Most cold outreach pours everything into the first email and treats follow-ups as a chore. The data says the opposite: 58% of positive replies land on follow-ups two through four. Here's how I rebuilt outbound-engine around the ladder.

My AI agent sent 200 cold emails perfectly. I never saw the replies.
One of my AI agents launched a cold-email campaign that sent perfectly and told me nothing — because it went out without a webhook. Here's the silent failure, the hard rule I made, and three more lessons my deployed agents taught me, all shipped in @di-atomic/smartlead v1.9.0.
Get new posts and release notes in your feed reader