albedo
small gleam coding and persistent agent daemon with a detachable cli and a canvas ui forked from okami.mom/albedo
Star 0
Fork 0
Watch

Version history of test/e2e/warm_test.py

October 2026

ptr.pet
daemon: children with bg jobs lets warm extension keep parent warm
committed
ptr.pet
warm: keep warming a parent waiting on children after its kernel is released
committed
ptr.pet
warm: rebuild the last turn call so a session still waiting keeps warming after a restart
committed
nekomimi.pet
warm test: 3s ttl so a loaded machine's tick lateness stays inside the drop window
committed
nekomimi.pet
test(e2e): hold kernel_upgrade's live job on a gate, not a 4s sleep
committed
nekomimi.pet
feat(warm): keep the cache warm for background jobs, on by default
committed
aly.codes
feat(api)!: migrate daemon and clients to protocol 3
committed
nekomimi.pet
test: share the daemon where a test only needs session isolation, and say why the rest cannot
committed

September 2026

ptr.pet
e2e: bound warm pings from the turn's send, the clock the warmer times them by
committed
ptr.pet
test: run every suite of the gate at once from one daemon snapshot, and shorten the quota and warm clocks
committed
ptr.pet
e2e: give each exclusive test a fresh daemon beside the shared one, keep test daemons offline, and reuse connections
committed
aly.codes
fix: reformat repo with treefmt
committed
ptr.pet
fade the footer's cached count as the prompt cache expires
committed
ptr.pet
Stop warming one ping sooner: floor(write / read) - 1
committed
ptr.pet
Move the prompt-cache warmer into a warm extension
committed
ptr.pet
Leave prompt-cache warming off until explicitly enabled
committed
ptr.pet
Call the per-attempt provider log requests, not a ledger
committed
ptr.pet
Keep an idle orchestrator's prompt cache warm while its children run
committed