-
Notifications
You must be signed in to change notification settings - Fork 0
OpenAug 19, 2026
No due date
•Last updated Three findings were drawn from single-attempt runs before 14 August, which Loop 1 showed cannot separate a real result from variance. They are not retracted and not confirmed. This sits after milestone 2 because re-running them against untrustworthy scorers would produce a second set of suspect conclusions.
25% complete
List view
0 of 3 selected 0 issues of 3 selected
Why do our skills make the weak model worse?
findingA result we do not yet understand and should investigateA result we do not yet understand and should investigateStatus: Open.#2 In hookdeck/evals;Confirm the docs-suppression mechanism rather than inferring it
findingA result we do not yet understand and should investigateA result we do not yet understand and should investigateStatus: Open.#10 In hookdeck/evals;Does documenting hookdeck ci stop agents giving up on CLI authentication?
findingA result we do not yet understand and should investigateA result we do not yet understand and should investigateStatus: Open.#27 In hookdeck/evals;