Xiangzhe e7858d7165 feat(video): cap the drill-down cache with a global byte budget ...

The per-session drill-down cache now tracks decoded bytes per entry and
evicts least-recently-used entries until an aggregate maxTotalBytes
budget fits (route sets 256 MiB); an entry larger than the whole budget
is rejected. Prevents the previous worst case of 64 x 32 MiB (~2 GiB)
pinned in memory.

2026-08-18 08:25:15 -03:00

..

_helpers

…

_setup

…

benchmarks

…

boundary

…

e2e

fix(providers): validate bailian-coding-plan against the Token Plan host (#10634)

2026-08-18 05:51:34 -03:00

fixtures

feat(providers): add local ZCode ACP backend (#10184)

2026-08-16 00:13:36 -03:00

golden-set

…

helpers

fix(sse): close the synthetic keepalive reasoning item + harden output_index allocation (#10330)

2026-08-16 00:15:38 -03:00

homolog

…

integration

fix(release): align agent skills catalog tests

2026-08-17 05:18:57 -03:00

live

…

llm-security

…

load

…

manual

…

security

…

snapshots

fix(providers): repoint freeaiapikey to its live API host and resync its catalog (#10233)

2026-08-16 00:42:25 -03:00

translator

…

unit

feat(video): cap the drill-down cache with a global byte budget

2026-08-18 08:25:15 -03:00

_cp_mock_hook.mts

…

_cp_mock_module.mts

…

_run_dns_guard_test.sh

…

e2e-tool-calls.sh

…

theoldllm-stress.test.ts

…