LoopFlow
Tutorial Workshop Keywords ๐ŸŽฎ LoopFlow Lab
Keywords / run / then run

run / then run

A step inside a flow. Compose

Syntax

run "<file.loop>"
then run "<file.loop>"
โ€ฆ with the result of <step>

What it does

A flow step. run "x.loop" runs every definition in that file; then run chains the next and receives the previous step's text summary; with the result of <step> pulls the handoff from a named earlier step instead of the previous one.

Where a loop self-corrects on one goal and a pipeline sequences stages inside a single file, a flow composes at the file level: each run executes a whole .loop โ€” its full planโ†’actโ†’observe cycle, its own done when โ€” and passes a short text summary forward as context for the next step. Reach for it when a job spans units you already built and tested separately: a build loop, then a test loop, then a deploy loop. The chain is fail-fast โ€” a step that ends unsatisfied halts the rest, so a broken build never reaches deploy. By default each step auto-carries the previous step's summary; with the result of <step> overrides that when a later step needs an earlier output instead (release notes that reference the build, not the test that ran after it). Paths are relative to the flow file. Add a human approves first under a step to gate it โ€” the flow blocks until a person signs off before that file runs, which is how you put a hand on irreversible steps like a deploy.

Example

flow "release":
  run "build.loop"
  then run "test.loop"
  then run "notes.loop" with the result of build
named handoff

How it runs

summary โ†’summary โ†’ ๐Ÿ“„ build.loopall loops inside ๐Ÿ“„ test.loopall loops inside ๐Ÿ‘ค approves first๐Ÿ“„ deploy.loopall loops inside runs in order ยท โœ— a file that isn't satisfied halts the rest
A flow chains separate .loop files; after each, a short text summary carries into the next as context.

Common mistakes

Related