AI skills & agents · Catalogue

A map of the work, not a list of skills

1,202 skills and agents, organised by where they bite in your practice: 5 practice models, 37process stages, granular down to the actual job. The library isn’t the offering. The offering is composing these into a skill set built for the way you already work.

1,202
Skills + agents
5
Practice models
37
Process stages

Three primitives, two surfaced

Agentic automation is built from three parts. Tools are the actions a system can take — call an API, query a database, drive a browser; the commodity substrate. Skills are packaged know-how that tells an agent how to do a specific job well. Agents decide which actions to take, in what order, toward a goal.

This catalogue is the inventory. The harder part is turning these parts into a skill set that fits one person’s practice, which is the AI skills & agents work itself. If you’re evaluating software solutions rather than AI skills, the solution-shapes library is the other catalogue.

Practice model

Authority-platform scholar-practitioner

260
skills + agents · 7 stages

A solo expert whose name and ideas are the product — speaking, writing, the platform built on a reputation.

Practice model

Independent advisor / fractional expert

187
skills + agents · 6 stages

One person selling embedded judgement over time — fractional roles, retained advisory, executive coaching.

Practice model

Boutique consultancy

262
skills + agents · 8 stages

A small firm running team-delivered client projects, where the principal coordinates as well as produces.

Practice model

Productized-expert / educator

272
skills + agents · 8 stages

An expert packaging what they know into courses, cohorts, and communities sold to many buyers at once.

Practice model

Expert-services micro-firm

221
skills + agents · 8 stages

A small firm running one repeatable expert deliverable as a service line — audits, diagnostics, intelligence reports.