Repository navigation
fix(quota): bound blocked-frontier Vision baseline initialization - #8574
xiaobaifly7 wants to merge 1103 commits into
Conversation
loopx-agent
left a comment
There was a problem hiding this comment.
Reviewer: model_agent · gpt-6.1-sol · OpenAI · runtime_reported · reasoning_effort=max
评审提交 2938c338851793a4e487bf4381ce7b8b10a4cb0d。结论 REQUEST_CHANGES。
P2 等待回执在重试时绕过一次性初始化限制
等待回执绑定了重规划身份。用同一个等待 Turn ID 重读后,回执升级,返回 should_run=true、must_attempt=true 和结算计划。 从原始正向准入回执派生豁免。安静等待应保留诊断义务,避免写入能在重试时绕过限制的结算身份。
P2 准入回执写失败仍消耗唯一初始化机会
只在真实 CLI 的 rollout 追加点注入 OSError。首次 guard 返回 heartbeat_receipt_write_failed,磁盘上已有准入 sidecar。修复 IO 后,新的 Turn 返回 agent_scope_wait。仍持有失败 Turn ID 的调用方可以原样重试并成功准入。 在现有回执提交边界确认一次性初始化,或把预留与真实回执对账,在明确失败后恢复资格。
动机 / Motivation
处理阻塞 Todo 的 Agent 需要一次有明确结算身份的 Vision 基线初始化,并能在后续唤醒进入安静等待。 自治重规划只剩缺失的 Vision 基线时,主干会为没有可执行 Todo 的 Agent 再次返回 run_now。 候选首次等待是安静的。两个新的 Turn 各自在相同身份下重试后都重新取得执行权,正常重试可以绕过单次初始化上限。 检查覆盖真实 CLI、回执和调度投影,执行范围到启动判定为止。
改动思路 / Approach
复用现有 agent_scope_wait,保留原始初始化身份,让等待和回执协调读取同一个有效准入事实。 新增本地准入记录需要与已有 heartbeat receipt 的真实状态一致,现有回执协调负责解释重试资格。 设计问题 #8572
具体改动 / Changes
10 文件,+299/-3。生产 +110/-2,测试和 fixture +189/-1。 新默认策略仍需 maintainer 决定。NEW-ADMISSION 的首次决策通过。ORIGINAL-RECEIPT 的同身份接续在下述情形失败。OWNER-STOP 仍走既有前置检查。主干基线 beff0f3e55d1f70eceebbd8e78c11f1cff34643b。docs/quota-allocation.md
build_quota_should_run。把结算绑定的存在直接转成 receipt_bound,进入 TypeScript 的豁免分支。 loopx/control_plane/quota/should_run.pyboundVisionBaselineAdmission。持锁写入一次性记录,并对后续新 Turn 投影安静等待。 loopx/control_plane/quota/vision_baseline_admission.tsquota_rollout_settlement_binding。结算计划为空时仍从诊断义务推导绑定,等待记录因此带上重规划身份。 loopx/control_plane/quota/settlement_cli.py
对主干的风险 / Risks
额度 1/0,原 Turn 已获得正向回执。另开两个 Turn,各自第一次查询均安静等待,使用各自同一 ID 再读 guard 后均升级成正向重规划。 只在真实 CLI 的 rollout 追加点注入 OSError。首次 guard 返回 heartbeat_receipt_write_failed,磁盘上已有准入 sidecar。修复 IO 后,新的 Turn 返回 agent_scope_wait。仍持有失败 Turn ID 的调用方可以原样重试并成功准入。
首次真实准入返回 autonomous_replan_required,首次新 Turn 安静等待。独立任务和到期监控的 File、SQLite 对照通过。 Python 109 项、TypeScript 91 项、控制面 typecheck、Ruff、quick premerge 6 项直接检查及 3 项选定检查均通过。
合成数据通过真实 CLI 和 TypeScript Effect runtime,分别用 File 与 SQLite Todo authority 验证。安装后的宿主和远端 CI 尚未复验。
我的整体评价 / Verdict
从原始正向准入回执派生豁免。安静等待应保留诊断义务,避免写入能在重试时绕过限制的结算身份。 用同一夹具验证原身份、新 Turn 和重复读取的结果。另加回执追加故障并验证修复后的重新准入。
English verdict: REQUEST_CHANGES for 2938c338851793a4e487bf4381ce7b8b10a4cb0d. A quiet wait receipt becomes executable on same-Turn replay. A failed guard append also consumes the sole initialization admission.
| prepared.vision_initialization_context = { | ||
| "runtime_root": str(runtime_root), "goal_id": safe_goal_id, | ||
| "agent_id": prepared.agent_frontier_id, "turn_instance_id": turn_instance_id, | ||
| "receipt_bound": bool(receipt_bound_todo_id or receipt_bound_replan_obligation_id), |
There was a problem hiding this comment.
[P2] 等待回执在重试时绕过一次性初始化限制
等待回执绑定了重规划身份。用同一个等待 Turn ID 重读后,回执升级,返回 should_run=true、must_attempt=true 和结算计划。 从原始正向准入回执派生豁免。安静等待应保留诊断义务,避免写入能在重试时绕过限制的结算身份。
| if (!previous) { | ||
| const admission: VisionBaselineAdmission = {schema_version: BASELINE_ADMISSION_SCHEMA, ...scope, | ||
| turn_instance_id: turn, replan_obligation_id: requireNonEmptyString(obligation.obligation_id, "obligation_id")}; | ||
| await durableWriteJson(path, admission); |
There was a problem hiding this comment.
[P2] 准入回执写失败仍消耗唯一初始化机会
只在真实 CLI 的 rollout 追加点注入 OSError。首次 guard 返回 heartbeat_receipt_write_failed,磁盘上已有准入 sidecar。修复 IO 后,新的 Turn 返回 agent_scope_wait。仍持有失败 Turn ID 的调用方可以原样重试并成功准入。 在现有回执提交边界确认一次性初始化,或把预留与真实回执对账,在明确失败后恢复资格。
|
This pull request has merge conflicts with Choose the remote for the base repository, not an out-of-date fork. git fetch upstream
git rebase upstream/main
# Resolve each conflict, git add the resolved files, then git rebase --continue.
git push --force-with-lease origin HEADFor a same-repository clone whose Keep the DCO |
Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…ive-readiness-recovery-20261119 Recover interrupted Codex readiness from the original Turn
…ch-case-reading-20261118t0349 Preserve saved case links when a research version cannot open
…eduler-fallback-20261119 Use one C1 scheduler proposal through Codex fallback
…oject#8790) * feat(research): capture authenticated LoopX attention receipts Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * fix(research): tighten attention receipt provenance Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> * fix(research): publish attention captures after source validation Signed-off-by: huangruiteng <huangrt01@163.com> --------- Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Signed-off-by: huangruiteng <huangrt01@163.com> Co-authored-by: TRAE CLI <traecli@bytedance.com> Co-authored-by: huangruiteng <huangrt01@163.com>
Signed-off-by: huangruiteng <huangrt01@163.com>
…sing in steward chat loops (loopx-project#9041) * perf(lark,native_chat): avoid full directory reparsing in steward chat loops - Make LarkOwnerAttention._route iterate capture candidates lazily so matching stops at the newest valid p2p capture. - Reuse FileChatDiscovery in PrivateIngressObserver._records while preserving strict field validation and fail-closed errors. - Reuse FileChatDiscovery in ChatExternalConversations.pending so unchanged request journals skip redundant JSON reads and stay mutation-isolated. - Add regression and scan-efficiency tests in tests/test_lark_steward_directory_scan_perf.py. Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(native-chat): project small recovery metadata fields to stop data_url memory growth Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…ct#9073) Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Co-authored-by: TRAE CLI <traecli@bytedance.com>
…ith handoff and user actions (loopx-project#9074) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…ject#9072) Remove redundant defensive prose while preserving published heartbeat wrappers and externally managed skill readiness. Validated with 197 related tests, independent historical wrapper migrations and negative cases, installed-skill readback, and current-main quick canary. Existing CI qualification and infrastructure limitations remain recorded in PR loopx-project#9072. Signed-off-by: lusendong.6789 <lusendong.6789@bytedance.com> Signed-off-by: Lusen Dong <17559402+cocolord@users.noreply.github.com>
…dit-chat-retry-feedback-8729-20261119 Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
… of mid-turn failed reports (loopx-project#9077) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…ct#9072 Restore non-redundant prompt-injection, non-authority, disclosure, and review-contract prohibitions alongside broken prose fragments, skipped invariant lists, and docstring regressions introduced in PR loopx-project#9072 (tracked in post-merge audit review 5976371492). - Restore prompt-injection and non-authority prohibitions in loopx/chat_manager.py, loopx/chat/coordination.py, loopx/chat_agent.py, loopx/presentation/answer_instruction.py, loopx/chat/manager_details.py, AGENTS.md (kept at 23,637 bytes <= 24,000), skills/, explore READMEs, manager_context skill, and review_contract.py (preserving policy_revision=26). - Restore dropped test assertions in tests/capabilities/test_pr_review_contract.py and tests/test_chat_intake_evaluation.py. - Repair broken prose fragments, dangling semicolons, lowercase continuations, skipped numbered lists, and orphan docstring lines across Dev Book, control-plane-course, guides, capability READMEs, and steward_turn.py / conversation_bindings.py. Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…onfirmation (loopx-project#9081) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…ker boundaries (loopx-project#9083) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…roject#8961) * refactor(content-ops): slim surface module and eliminate Any Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(content-ops): reject malformed angle source_item_ids containers Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…Goal (loopx-project#9085) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…oopx-project#9087) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…cope runtime revision in handoffs query (loopx-project#9088) Eliminate the remaining O(N) Effect runtime RPCs and uncached runtime fingerprinting inside tracking.query(): enter effect_runtime_request_scope() for the query, memoize pure history_inspect lifecycle decisions across entries sharing identical Goal refs while holding the Goal lifetime guard, lazily discover legacy ingress files only when external audience recovery needs them, and preserve entry scan order across Goal batches. Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…LOC reduction (loopx-project#8929) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…led releases and extend readiness timeout to 90s (loopx-project#9089) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…opx-project#8885) * refactor(control-plane): split and slim quota monitor_poll_commit Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(control-plane): preserve legacy lane budget and observation diagnostic order Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…oopx-project#8776) * fix(turn): refuse a run-once preparation claim that was not consumed When a should-run decision advertises a managed claim before work and the builtin run-once bridge declines it, the flow previously fell through to the generic not-host-executable rejection. A bounded caller then retried the identical blocked route without ever invoking the Host. Refuse with a typed boundary instead: the run-once packet now carries status=execution_preparation_required with the turn_execution_preparation _claim_not_consumed reason code, the advertised claim command and the bound Todo, and exits without host invocation or quota spend. The existing consumption path, its admission wait and the ready postcondition are unchanged. Signed-off-by: BigDataDZ <76271875+BigDataDZ@users.noreply.github.com> * Keep run-once preparation refusal scoped to Codex CLI Leave native App Server preparation on its existing path. Render a typed execution refusal with the verified claim and readiness requirement when the builtin CLI cannot prepare work. Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: BigDataDZ <76271875+BigDataDZ@users.noreply.github.com> Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> Co-authored-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
* docs(stride): define meta-goals within a Todo Signed-off-by: cicolord <cicolord@outlook.com> * docs(stride): define meta-goals and scheduler acceptance Signed-off-by: cicolord <cicolord@outlook.com> * docs(stride): compare feasible continuations by expected value Signed-off-by: cicolord <cicolord@outlook.com> * docs(stride): work through the scheduler probe decision Signed-off-by: cicolord <cicolord@outlook.com> * docs(stride): estimate and calibrate scheduler decisions Signed-off-by: cicolord <cicolord@outlook.com> --------- Signed-off-by: cicolord <cicolord@outlook.com>
…nd bucket wakeups (loopx-project#9207) * feat(steward): refine stalled PR signal clock, author-reply bypass, and bucket wakeups Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(steward): scope stalled-age hash suppression, under_24h bucket, and maintainer reply check Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(steward): paginate stalled-PR comment suffix and preserve incomplete coverage Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(steward): gate comment backfill to age monitors and isolate non-age hashes Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…ease pipeline (loopx-project#9244) * Upgrade multi-surface release preflight, DSH npmjs discovery, and release pipeline Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Address PR 9244 review findings across preflight, verifier, and docs Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…ler-observation-recovery-gate-20261122 control_plane: project prior-turn recovery gate on --scheduler-observation
…laim-ttl-ergonomics-20261122 control_plane: add bounded --ttl-seconds on todo claim and clarify source.ref diagnostics
…oopx-project#9021) * fix(reports): retain verified terminal failure evidence Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(reports): durably invalidate contradicted report failure proofs Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(steward): separate unavailable items from contradictions and neutralize failed writes Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(steward): persist durable parent contradiction before attempt cleanup Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…nch-cli-catalog-fix Fix Workbench command catalog and manual
…a/b release gate Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…oopx-project#9250) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…behind a symlinked path Account discovery decided whether an instance already claimed an account by comparing CLAUDE_CONFIG_DIR strings. When the daemon's HOME is spelled /data00/home/x while the registered instance uses /home/x (a symlink to the same directory), the account looked unclaimed and auto-join added it again (e.g. claude-4 / claude-5 duplicating claude-2 / claude-3). The removed-account blacklist had the same string comparison, so a removed duplicate could come back. Compare absolute env paths by real path in both places. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Signed-off-by: maxliux5 <5958865+maxliux5@users.noreply.github.com>
…-commander-evaluation-20261122 feat(release): add four-state Global Commander evaluation ledger and a/b release gate
Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…f987 fix(workbench): stop auto-join duplicating Claude Code accounts behind a symlinked path
) * Recover Steward steering receipts after Chat restart Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Label queued steering recovery as checking the original insertion Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> Co-authored-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…oopx-project#9222) * fix(collaboration): verify terminal return receipt identity Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(collaboration): reconcile private terminal returns from saved messages Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Read detailed delegation returns from the Inbox projection Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> Co-authored-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…t_pending Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…72-tail-20261121 fix(docs): repair remaining prose definitions, invariant titles, and comment fragments from loopx-project#9072
…spend-index-conflict-20261121 Retry a definitively rejected Turn quota spend
…ject#9227) * Show when queued input changes require the original source Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * Keep bound queue drafts readable without blocking steering Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> * fix(chat): explain restricted queue actions and retain closed drafts Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com> --------- Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
…d prep (loopx-project#9265) Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
Signed-off-by: LoopX Agent <337587101+loopx-agent@users.noreply.github.com>
loopx-agent
left a comment
There was a problem hiding this comment.
Reviewer: model_agent · gpt-6.2-sol · OpenAI · runtime_reported · xhigh
动机 / Motivation
仅有受阻事项的 Agent 需要保留原 Vision 诊断,同时避免每轮反复创建初始化任务。 主干在 File 与 SQLite 的独立进程对照里,把四次守卫全判为自主重规划,得到四个不同 Turn 绑定。 有可办任务、到期 Monitor、其他 Agent 或其他 Goal 沿用原独立执行规则。被阻塞的同一范围仅限制缺失基线所引起的重复初始化。 候选将唯一一次批准绑定到成功提交的正向回执。竞争失败者保留诊断和安静等待。 改动在缺失基线的阻塞前沿、回执持久化与原 Turn 的重放。
改动思路 / Approach
TypeScript 的 quota owner 给出类型化值,使用原 rollout append lock 串联实际提交。 按当前 Goal lifetime、Agent 和原 Turn 区分,相同身份的未完成结算保持原跟进。 沿用现有 rollout log 的提交条件。等待只携带诊断,正常激活和原 Turn 同一绑定继续走原 owner。
具体改动 / Changes
准确头 e94e81e,merge base c563a47,15 个文件,+584/-10。按实际远端主干逐项检查原 contributor 与后继修复。核验初始化票据、并发、原结算和已有标准等待。 按修订 c563a47 的现有规范核对各项要求。 目标规范
关键代码讲解
-
boundVisionBaselineAdmission基于完整工作计数和已提交的正向 receipt。 规则 -
_check_vision_baseline_admission在追加锁内检查同一范围的回执。 追加 -
quota_rollout_settlement_binding让等待诊断回执保持无绑定。 绑定 -
普通 CLI 回归模拟追加前后错误和竞争者。 回归
-
initialization_receipt核验换行提交和竞争。 -
original_turn核验赢家和等待者重放。 -
ordinary_scope核验独立工作和到期 Monitor。
对主干的风险 / Risks
两种存储各用四个真实并行 CLI 进程,在同一 Goal 与 Agent 上竞争初次初始化。主干每种存储均准入四个独立 Turn。候选均只有一个准入、三个等待。再次调用赢家保持原结算身份,再次调用等待者仍无结算绑定,预检查未产生扣额。 在 File 和 SQLite 上注入追加前失败、缺提交换行以及提交后丢失响应。前两种恢复后重新准入,提交后的重试恢复同一原身份。新可办工作或到期 Monitor 保留正常执行分支。 定向 130 项 Python 和 106 项 TypeScript 通过。独立并发对照每侧各覆盖 File 与 SQLite,TypeScript 类型检查和 Ruff 通过。 另外重验原票据、回放与旧等待场景,保留实测结果 126 passed。 GitHub PR 的 base 快照仍显示旧引用。以经 GitHub compare 与本地 Git 双重确认的当前 main 为评审基准,其净差异为 15 个文件。主干随后合入 #9235,六处更新与本提案路径无交集,已核对三方合并。标准预合并的 CLI output budget 检查触发 120 秒外层时限,原标准结果仍失败。 独立运行同一命令,主干 311 秒与候选 292 秒均完成,通过内容与大小断言。 签名安装、远端 CI 和 PostgreSQL 在本轮未独立验证。
我的整体评价 / Verdict
批准当前提交,未发现阻断项。原两项求改所列故障已在当前提交通过独立回归。代码结论限当前主干上的上述 15 处改动。 CLI guard 和结算语义改变,当前 diff 无首屏渲染变化。
English verdict: APPROVE at e94e81e. Approve the 15-file source delta against the verified current main. Four independent guards converge to one committed admission on both File and SQLite. Failure before commit preserves an opportunity to retry, and waiting receipts cannot bind a replan on replay. The outer premerge CLI budget result and GitHub base comparison discrepancy are recorded separately.
The merge-base changed after approval.
Goal And Delivered Outcome
agent_scope_waitcontract while preserving the missing baseline obligation. The same synthetic input on pinned main returnedautonomous_replan_requiredagain on a fresh Turn.main. This is standalone, not stacked on fix(quota): wait when the only scoped work is blocked #8511.Author Declaration
Implemented against
agent_scope_frontier_v1/agent_scope_waitbehavior is reused.quota/vision_baseline_admission.ts::boundVisionBaselineAdmissionagent_scope_frontier, existing interaction/scheduler projectiontest_missing_vision_baseline_wait.py::assert_waitchecks no attempt/delivery, quiet wait and scheduler backoffScope And Continuation
POLICY CHOICES
agent_scope_wait/ scheduler backoff while retaining the pending Vision obligation and the original admission identity in the frontier. No quota debit or successful-baseline receipt is fabricated.These are proposed conservative defaults, not a claim of a prior maintainer ruling. Maintainers may change the allowance, consumption point, scope or rearm rule.
Validation
build_quota_should_runwith disposable runtime and real File/SQLite authority; fresh Python process reads the durable admission and waitsautonomous_replan_required; same synthetic scenario on this branch waits after initializationPYTHONUTF8=1passes without a source/config changeLOOPX_USAGE_PING=0; Windows commands also usePYTHONUTF8=1):Frontend / Visual Evidence
Type of Change
LoopX Area
Technical Direction
Shared-authority RFC fixture impact
Boundary Checklist
Maintainer update
The updated branch integrates current
mainand preserves the original contributor commit.agent_scope_waitreceipt retains the pending Vision diagnosis. It carries no Todo or replan settlement binding, so replay of the same Turn stays quiet.Exact PR head
e94e81ea7310e3ec29992b685c523e661d720895. The focused suites passed 82 Python and 106 TypeScript cases. Typecheck, Ruff and the public-boundary scan passed. The final change after those runs preserves the Python AST and meets the changed-file line ceiling. Exact-head standard premerge is running. Independent review remains pending.