Source-first workflow

Michael says coding agents learn a fast-moving library more reliably from its source repository. Stale model knowledge and large prompts are less reliable. 08:51 12:59

He describes compiler diagnostics, tests, lint checks, and saved patterns as feedback for the next move.

Diagnostics report problems found by the compiler. Lint checks reject code patterns that break project rules. He says patterns should fit the project and model. 30:30 58:08

See Vibe engineering Effect apps6.

Sources

  1. Vibe Engineering Effect Apps — Michael Arnaldi, Effectful - YouTube. Michael Arnaldi. Source-first demonstration; used for learning a fast-moving library from its repository. Accessed 2026-10-01.
  2. Vibe Engineering Effect Apps — Michael Arnaldi, Effectful - YouTube. Michael Arnaldi. Workshop explanation; used for the limits of stale model knowledge and large prompts. Accessed 2026-10-01.

Last changed 2026-10-07 in e73c9d9. Source on GitHub. Change log.

Linked from

  1. Agent guide. An Effect stack so pure (aspirational) Kit Langton will blush.

    Read the source repository and gather the fields below.

  2. Change log. What changed in the files served here, newest first.

    Source-first workflow — edited: feat(learn): teach the next twenty cards with plain lines, diagrams and snippets

  3. Full agent guide. An Effect stack so pure (aspirational) Kit Langton will blush.
  4. Rat Stack lore | rat-stack. Short, source-grounded notes on the ideas and decisions behind rat-stack.

    Source-first workflow — Michael Arnaldi says agents should inspect a library’s current source instead of trusting stale model knowledge.

  5. Source change history (log.md). The generated change log as Markdown.

    Source-first workflow — edited: feat(learn): teach the next twenty cards with plain lines, diagrams and snippets