Clean Session Handoff
End a coding session in a state the next session (or agent) can pick up without re-discovering context.
CATALOG NO. 001 · JULY 2026 EDITION
Field-tested loops, goals, workflows, and prompts. Look one up, copy it, paste it into Claude Code, Cursor, or Codex. No account, no theory, no fluff.
30 entries in the catalog
End a coding session in a state the next session (or agent) can pick up without re-discovering context.
A six-step loop for any coding task: clarify, check, choose, execute, verify, recommend.
Make the agent build, run, diagnose, and retry until there's real passing evidence, not a hunch.
Force one hypothesis at a time against real evidence, so the AI stops guess-and-check fixing.
Score AI-generated code against a checklist and keep iterating until every item passes.
Don't accept 'fixed it' from an AI agent until it re-runs the exact failing case and shows real output.
Run before closing any coding session so nothing sits uncommitted overnight.
Confirm build, routes, and diff are clean before you push to production.
Write a handoff so the next session or teammate starts working in minutes, not re-explaining.
Classify a task's complexity first, then gate execution behind your approval so agents stop over- or under-building.
Feed a one-line idea in, get a real PRD out: problem, goals, non-goals, stories, P0/P1/P2.
Expand a one-liner into success criteria and a verify step so the agent can loop without you babysitting it.
Turn messy notes into a publish-ready blog post or landing page section in one pass.
Turn a short project brief into a structured client proposal or RFP response draft.
Edit AI-generated copy to cut filler, hedging, and generic AI phrasing without losing meaning.
Give the agent a checklist of AI-slop tells and make it revise its own UI until it passes.
Feed it three colors and a vibe, get back a concrete palette, type scale, and spacing scale to reuse everywhere.
Force your AI to describe the mockup and wait for your yes before touching production UI.
Build a clean llms.txt index so AI answer engines cite your site accurately.
Paste a page's content, get back valid JSON-LD structured data ready to embed.
Audit a landing page's indexability and long-tail targeting, get a prioritized fix list.
Map licensing, compliance, and recent policy shifts for a product launch in one structured pass.
Scope TAM/SAM/SOM for a product idea in one pass, with every number labeled estimate or sourced.
Turn a rough audience description into grounded user segments with needs, objections, and buying triggers.
Turn a competitor's public pages into a structured pricing and feature comparison table.
A weekly check that flags only what changed on a competitor's public pages, not the whole page again.
Compare your product against competitor URLs to find the gaps worth leading with in messaging.
Expand a one-line campaign idea into a full brief an agent or agency can execute against.
Paste your changelog entry, get a structured hook-problem-solution-proof-CTA launch thread.
Feed an agent your weekly metrics and get one ranked growth lever with the reasoning shown, not five vague ideas.