Here are
6 public repositories
matching this topic...
RLAR as an hmz flow: an actor works in one session until a fresh reviewer, reading the repository each round, says the task is done.
Updated
Nov 18, 2026
Python
The harness's own /goal as an hmz flow: the task set once as the agent's goal, worked until it is met or the budget is spent.
Updated
Nov 18, 2026
Python
The continue loop as an hmz flow: the task once, then "continue" in the same session until the budget is spent.
Updated
Nov 18, 2026
Python
The Ralph loop as an hmz flow: the task again and again, a fresh session every round, until the budget is spent.
Updated
Nov 18, 2026
Python
The flame chase as an hmz flow: two agents take turns on the same task, each turn a fresh session, until the budget is spent.
Updated
Nov 18, 2026
Python
The Ralph loop with memory, as an hmz flow: the task again and again in one session that remembers every round.
Updated
Nov 18, 2026
Python
Add this topic to your repo
To associate your repository with the
flowbench
topic, visit your repo's landing page and select "manage topics."
Learn more
You can’t perform that action at this time.