weichen/Agent-System-Backend
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Keep original materials and cleared results reachable across turns (quality P2-I1) #112

Merged
weichen merged 3 commits from feature/agent-quality-phase02 into main 2026-10-09 15:22:46 +00:00
Conversation 0 Commits 3 Files Changed 23
+339 -58
3 Commits
Author SHA1 Message Date
weichen 9fe2a592f3 Merge origin/main into feature/agent-quality-phase02
ci / vet-and-test (pull_request) Successful in 43s
Details
2026-10-09 15:12:16 +00:00
weichen 908fb253d7 Keep tool-result dumps out of the deliverable gate and cards
ci / vet-and-test (pull_request) Successful in 45s
Details 
Why: The live after-run (golden minutes-pm-tool) showed the truncation spill
artifact tool-run_shell-<call>.txt entering CheckArtifacts as a .txt
deliverable; its non-UTF-8 payload failed the scorer four times in a row.
These files exist to keep oversized tool output reachable - they are context
relief, not deliverables, and must not occupy delivery cards either.

Evidence: SELF - TestLatestDeliverablesKeepsLastPerName extended with a
tool-dump ref and asserts exclusion; go test ./internal/... passes; single-task
live rerun recorded on the phase02 task card.

Impact: latestDeliverables (gate input, verifier refs, reply deliverable list,
final artifact_ref cards) skips kind=data tool- prefixed refs. The spill files
remain artifacts for context backfill.
 2026-10-09 14:56:34 +00:00
weichen 43297795f5 Keep original materials and cleared results reachable across turns
ci / vet-and-test (pull_request) Successful in 48s
Details 
Why: Real-run before/after evidence (phase01 card) showed later turns never
re-read the uploaded source (0 original reads in plan-followup) because the
attachment only existed in the current turn's injection and the context window
hard default was 32000 chars - a 25k-char transcript tripped the clear-on-first
request, replacing old tool results with a pointer the model has no tool to
follow. MiniMax is an interleaved-thinking model whose docs require <think>
blocks in assistant history; dropping them degrades later turns.

Evidence: SELF - new tests TestThinkReplayWrapsAssistantHistory,
TestContextWindowScalesFromTokens, TestThinkReplayEnabled,
TestPersistInputTextsAndListWorkspaceInputs; go test ./internal/... passes;
live before/after comparison on the golden task set recorded on the phase02
task card per the new policy clause.

Impact: models.chat.contextWindow/outputReserve/contextBuffer (tokens,
defaults 128k/16k/4k) drive the in-run clear threshold, converted to chars;
cleared tool results spill to <sandbox>/.agent/results with readable paths;
attachment parse text persists as inputs/<name>.md and the env section lists
all workspace inputs every turn; context.thinkReplay (auto|on|off, auto =
MiniMax) replays thinking blocks into request-side assistant history only.
 2026-10-09 14:28:39 +00:00