Mental models, operating principles, and repeatable AI practices
421 recursos públicos · 105 recursos revisados por la redacción
Actualizado: 2026-10-24T17:03:20.185066+00:00
A four-step progression that turns raw notes, pasted listings and safety sheets into reviewable briefs where observed facts, single-source claims, and human judgment are kept apart. Each step states what the prompt can and cannot do and ends with pass/fail checks you run yourself.
4 miembros de la colección
Actualizado: 2026-10-06T17:16:28.854199+00:00
A curated progression of four copy-paste prompts that transform unstructured text into auditable tables, strictly separating observation from interpretation and marking missing data explicitly.
4 miembros de la colección
Actualizado: 2026-10-10T10:16:19.695876+00:00
A four-step progression of copy-paste prompts that turn messy notes, listings, safety sheets and pantry lists into auditable tables. It strictly separates observation from interpretation, flags missing data, and forbids the assistant from acting or inventing facts.
4 miembros de la colección
Actualizado: 2026-10-19T00:16:38.961168+00:00
A four-step progression of reusable prompts that turn raw notes, listings, safety sheets, and pantry lists into structured, reviewable drafts. It strictly separates observation from interpretation, flags missing data explicitly, and forbids the assistant from acting or inventing facts.
4 miembros de la colección
Actualizado: 2026-10-19T16:16:55.872964+00:00
A four-step progression that transforms raw notes, listings, safety sheets, and pantry lists into auditable tables. It strictly separates observation from interpretation, flags missing data explicitly, and forbids the assistant from acting or inventing facts.
4 miembros de la colección
Actualizado: 2026-10-23T00:36:17.425377+00:00
A four-step progression that transforms raw notes, listings, safety sheets, and pantry lists into auditable tables. It strictly separates observation from interpretation, flags missing data explicitly, and forbids the assistant from acting or inventing facts.
4 miembros de la colección
Actualizado: 2026-10-23T17:16:23.921808+00:00
Ten picks for the dev building plan-then-execute agents (ReAct, ReWOO, Plan-and-Solve): planning frameworks, task decomposition, persistent plans, world-model agents, and the eval harness that proves your plan was actually better.
10 miembros de la colección
Actualizado: 2026-06-30T19:00:02+00:00
DeepEval, Promptfoo, Ragas, Opik, Guardrails AI — score every prompt change before it ships and catch regressions early.
5 miembros de la colección
Actualizado: 2026-06-30T19:00:02+00:00
Six GitHub-verified building blocks to ship safer agents: CI-grade eval harnesses, a GitHub Action gate, curated Cursor rules, and rule generation to keep guidance fresh.
6 miembros de la colección
Actualizado: 2026-05-11T23:08:06+00:00