skills library

Runs a full client AI-implementation engagement end to end. Seven phases, 33 steps, from first interview to a graded, deployed playbook. It's the skill that calls the other skills at the right moment.

Scores every candidate process on value vs. complexity, sorts it into build-now, defer, or cut, and picks the right Claude surface before a single line gets written.

Interviews the person who already does the process well, strips every input down to only what actually changes run to run, and checks every step against one rule: could a stranger follow it without asking a question.

Turns a design into a running system: project structure, an operating manual, tested tools, scheduled Routines, and a debugging table for when a run fails.

Researches the week's AI news, visits my curation links and writes them up in my voice, then slots everything into the email template. I write the intro. It does the rest.

Diagnoses whether a system, project, or week is missing structure, energy, or consciousness before it burns someone out, and helps design new things with all three built in from the start.

Diagnoses drift in my own writing: borrowed authority, abstraction without an anchor, theory without my skin in it. Flags the pattern, names the principle it breaks, hands back two or three rewrites so I pick, not it.

Runs alongside a standard business audit and listens for time sinks, repetition, and bottlenecks, then scores each opportunity against an AI fluency framework — effective, efficient, ethical, safe — before it's ever proposed.

Takes raw input — a transcript, a form, a pile of notes — and structures it into a complete 5Ws document, then surfaces every playbook hiding inside that domain, each with its own trigger-to-output flowchart.

Turns scattered process knowledge into one chart: what the business does, who's really operating each area today, and where a founder is quietly the single point of failure.