✔ native seed framing reserves capacity before optional history without changing the selected window (2039.7228ms) ✔ direct code evidence remains selected while recent historical-index browsing is archived within the complete request budget (2446.9514ms) ✔ one long user turn becomes bounded complete groups with exact UTF-8 history and recent JSON arguments (3863.3558ms) ✔ tool definitions, exact system bytes, output reserve and remaining Run/Task budget are accounted before requests (928.2012ms) ✔ oversized required user instructions pause intact and can be explicitly resumed after increasing the window (1714.7245ms) ✔ all task instructions, constraints and acceptance remain in the runtime system instead of an eight-item or byte prefix (1350.4595ms) ✔ a newest manageable multi-tool group stays exact even beyond the working target; oversized earlier groups are archived (809.835ms) ✔ invalid orphan result is rejected before archive or network preparation (592.7567ms) ✔ invalid missing result is rejected before archive or network preparation (714.997ms) ✔ invalid wrong result is rejected before archive or network preparation (657.1018ms) ✔ invalid duplicate call is rejected before archive or network preparation (726.3918ms) ✔ a group larger than the artifact limit uses verified Unicode-safe chunks without replacing tool JSON (1934.7044ms) ✔ changed or missing archived evidence stops reuse and is never silently recreated (1353.31ms) ✔ transaction failure prevents publishing a plan and any corresponding model request (1167.6815ms) ✔ the Provider rejects body mutation after a sealed plan before calling HTTP (1295.8821ms) ✔ HTTP failure and explicit same-Task resume reuse exact completed results and request hashes without repeating a command (12256.14ms) ✔ a real process exit after plan persistence marks interruption and never issues an automatic model request (1493.1791ms) ✔ small code and actual failed command stay visible when their original multi-tool group cannot fit (2817.5421ms) ✔ mixed group retains six distinct small file observations with exact multilingual text and bounded request bytes (2935.5331ms) ✔ working observations are bound to the current Task and stop treating a pre-application file revision as current evidence (2209.6653ms) ✔ a sealed observation rejects changed or missing original result artifacts instead of recreating evidence (1974.3203ms) ✔ a real passed-check observation is excluded after a source edit invalidates the check, while its historical sent evidence remains valid (6319.2793ms) ✔ budgeted file pages reconstruct UTF-8 text including long single lines and surrogate boundaries (4.4512ms) ✔ actual HTTP multi-tool reads survive an oversized runtime snapshot and the sealed evidence survives reopening (18421.1531ms) ✔ actual HTTP file pagination advances over a long file with each returned page visible in the next request (2954.4126ms) ✔ later mixed history pages cannot displace the actual module sources and complete requirements (5429.4631ms) ✔ a later narrow page of the same file revision cannot evict the wider actual observation (3259.3701ms) ✔ six small source files stay complete across metadata turns even when mandatory evidence exceeds the soft target (2831.2465ms) ✔ historical read-artifact navigation points to the original artifact and next page, never its result wrapper (1134.8138ms) ✔ a 1.6.1 full-provenance sent view remains verifiable after introducing the compact view (1190.842ms) ✔ actual HTTP requests keep four observed modules after a later large history read and preserve all source bytes (9762.9155ms) ℹ tests 31 ℹ suites 0 ℹ pass 31 ℹ fail 0 ℹ cancelled 0 ℹ skipped 0 ℹ todo 0 ℹ duration_ms 61529.4006