Playbooks
Outcomes, not capabilities
Every playbook below is a named outcome backed by a real agent template: instructions, a starter skill set, and the integrations it expects. Clone one and a short interview personalizes it to your product, audience, and voice.
Publish consistently, in your own voice
Clone an agent that learns how you write from samples you already published, researches each topic from primary sources before drafting, and hands you a second-draft post or thread you are happy to ship - never one it publishes on its own.
Triage support without the backlog
Clone an agent that classifies every incoming thread by topic, severity, and sentiment, drafts a reply grounded only in your docs and past answers, and escalates billing, legal, and angry-customer threads to a human immediately instead of guessing.
Turn an open question into a decision-ready brief
Clone an agent that breaks a research question into sub-questions before gathering sources, extracts claims into structured notes with citations, actively seeks the strongest counter-case, and writes a brief with a recommendation and a stated confidence level.
Ship release notes from what actually merged
Clone an agent that regroups merged changes by user-facing impact instead of by commit, leads every entry with the benefit, puts breaking changes first with a migration step, and keeps a consistent changelog voice release to release.
Get Smarter About AI Dev
New tutorials, open-source projects, and deep dives on coding agents - delivered weekly.
One email per weekReal code, not theoryFree forever