Book · $24.99
Agent Skills Pack: 17 worked examples
What each skill produces: the request, and the real output
See what good agent output looks like before you ask for it. For each of 17 skills, the book prints the request, the output that came back and how it was produced, and says plainly what was not run.
Who it’s for
- Developers using Claude Code who want reviewable examples of tests, reviews, diagrams and docs.
- Anyone writing their own skills who wants to see well-specified ones next to their output.
What you get
- Engineering Testing strategy (a failing test first, then the fix), dev containers, CI checks, code review findings, release and rollback.
- Agent harnesses Harness setup, subagent briefs, agent evals, a hooks and permissions audit, cost budgets.
- Diagrams and docs Architecture diagrams, labeled exploded views, decision records, READMEs and runbooks.
- Learning content PDF guides, quizzes and interactive HTML explainers.
Terms
- Format
- One downloadable PDF, 70 pages including covers, November 2026, delivered by Gumroad right after checkout.
- Scope
- The PDF only. The code kits, installers, skill folders and templates the books describe are separate and not included.
Honest notes
- Independent publication by Tim Urista. Not affiliated with or endorsed by Anthropic. Claude and Claude Code are trademarks of Anthropic.
- Example projects and team numbers are synthetic and labeled as such. They are not customer results.
- No security, savings or productivity guarantees.
Inside the book
Actual rendered pages from the PDF, with notes.
-
Agent Skills Pack, pages 10 and 51: the request and what came back, side by side. -
Agent Skills Pack, page 52: a real output of the exploded-view skill.