diff --git a/docs/architecture/rfcs/agent-loop-effect-interpreter-v0.md b/docs/architecture/rfcs/agent-loop-effect-interpreter-v0.md index fbc730f28b..840b67f9df 100644 --- a/docs/architecture/rfcs/agent-loop-effect-interpreter-v0.md +++ b/docs/architecture/rfcs/agent-loop-effect-interpreter-v0.md @@ -576,7 +576,10 @@ changes no scheduler admission, execution grant, default or host effect. Real packaged transport qualification remains separate from live model outcomes, App adoption and the three convergence acceptances below. -Remaining implementation Todos, in dependency order: +The requirement-delivery (#5844), complete closeout-clause (#6337) and original +Host decision-capture (#6350) source repairs have merged. Preserve those results; +their completion does not close M7.4, R5 or the three acceptance streams below. +Continue in this order using the existing owners and checkpoints: | Todo | Observable outcome and decisive acceptance | |---|---| @@ -596,6 +599,28 @@ This closes the producer-to-detail transport gap only. Live models, packaged App/Lark, later authority checks and the remaining convergence matrix retain their independent acceptance; no Python projection is retired by this slice. +The first ordinary-CLI consolidation replaces an exactly identical +`heartbeat_recommendation.blocked_priority_fallback` mirror with +`blocked_priority_fallback_ref="$.blocked_priority_fallback"`. The complete root +fact, notification, execution obligation and original capture remain intact; +different mirrors are retained, including Boolean/number differences. The existing +`--include-detail agent-todos` view retains the legacy mirror. File/SQLite CLI +tests cover same-capture action dimensions and no-debit replay. Original source +hashes intentionally distinguish renderings; this does not change signature +coverage or make a rendered view a substitute for the original capture. +Legacy-field access stays in the existing packet owner; the CLI invokes its +projection function without adding another legacy migration consumer. + +Next, inventory authorized real Goal captures across active work, peer/replan, +wait/recovery and optional-capability shapes, then map each retained clause to +its shared typed owner and actual consumer. Migrate one complete caller family +before retiring its replaced projection. Compare the same captured decision +through full/thin heartbeat and Turn, counting required detail reads, tokens, +serialization/IO, latency and resource growth. Byte savings from this mirror +removal qualify only that rendering delta; model quality, complete costs, +installed App/Lark adoption and migration defaults still require their own +evidence. Keep private source material out of public fixtures and checkpoints. + The earlier cost slice reproduced a regression-budget gap, rather than a missing fixture alias or permission to remove required context. Comparing `aa87cc019` with `fc411c878` on the unchanged public CLI fixture, identical diff --git a/docs/architecture/rfcs/agent-loop-effect-interpreter-v0.zh-CN.md b/docs/architecture/rfcs/agent-loop-effect-interpreter-v0.zh-CN.md index e71a880f07..de13ceb6f1 100644 --- a/docs/architecture/rfcs/agent-loop-effect-interpreter-v0.zh-CN.md +++ b/docs/architecture/rfcs/agent-loop-effect-interpreter-v0.zh-CN.md @@ -394,7 +394,9 @@ TurnEnvelope 投影，不能只用未绑定的合成 argv。适配器只为递 不改变 scheduler 准入、执行授权、默认值或宿主效果；打包后的传输验证不代表 live 模型结果、App 采用或以下三项收敛验收完成。 -后续实施 Todo 按依赖顺序推进： +完整要求交付（#5844）、完整收尾条款（#6337）和原始 Host 决策捕获（#6350）的 +源码修复已合入。保留这些成果；它们的完成不关闭 M7.4、R5 或下列三条验收线。 +后续沿既有 owner 和 checkpoint 按此顺序推进： | Todo | 可观察结果与决定性验收 | |---|---| @@ -410,6 +412,22 @@ fixture。只读规划与可选 quota capture 默认值不变，原生执行 Tur 这只关闭 producer 到 detail 的传输缺口。真实模型、打包 App/Lark、后续 authority 检查 及其余收敛矩阵仍独立验收，本切片不退役任何 Python 投影。 +首个普通 CLI 整合将完全相同的 +`heartbeat_recommendation.blocked_priority_fallback` 副本替换为 +`blocked_priority_fallback_ref="$.blocked_priority_fallback"`。根字段的完整事实、 +通知、执行义务和原始 capture 保留；不同副本继续保留，包括 Boolean 与数字的差异。 +既有 `--include-detail agent-todos` 视图保留原副本。真实 File/SQLite CLI 覆盖同次 +捕获的动作维度及无扣费重放。原始 source hash 本就区分不同渲染；本批不改变签名 +覆盖，也不让渲染视图替代原始 capture。 +旧字段访问保留在既有 packet owner；CLI 调用其投影函数，不增加待迁移的旧字段消费者。 + +下一步先调查获准真实 Goal 的活跃工作、peer/replan、等待/恢复和可选能力形态， +将每条保留义务映射到共享 typed owner 与实际消费者，再完成一族调用方迁移，最后 +退役其旧投影。用同次捕获决策比较 full/thin heartbeat 与 Turn，将必需详情补读、 +token、序列化/IO、延迟及资源增长一并计入。此次移除副本的字节节省只证明该渲染 +差异；模型质量、完整成本、安装态 App/Lark 采用和迁移默认仍需独立证据。真实私有 +来源不进入公开 fixture 或 checkpoint。 + 此前成本切片已复现回归预算缺口，不能归因于 fixture alias，也不授权删掉 必需上下文。`aa87cc019` 与 `fc411c878` 使用相同公共 CLI fixture、临时 alias 和命令参数，实际 JSON stdout 测量如下： diff --git a/docs/reference/protocols/turn-envelope-v0.md b/docs/reference/protocols/turn-envelope-v0.md index f7595cf4b6..23f440043f 100644 --- a/docs/reference/protocols/turn-envelope-v0.md +++ b/docs/reference/protocols/turn-envelope-v0.md @@ -51,13 +51,29 @@ not verify file integrity or make saved admission current. Missing, malformed, or mismatched files require recovery, never automatic guard replay. Consumers already referring to `full_decision`, such as compacted capability context and peer diagnostics, now read this same observation. Scheduler-specific legacy detail requests are unchanged. Replan obligations and fresh selection/lease/quota checks retain their existing owners. -Without capture, the historical full-decision route and all default outputs are -unchanged. This opt-in detail link does not change action-signature coverage. +Without capture, the historical full-decision route remains available. Enabling +capture does not enable TurnEnvelope or change admission/settlement defaults. +This opt-in detail link does not change action-signature coverage. 中文：同时开启短包和完整 decision 保存时，详情入口读取本次保存的文件，不再 为补读上下文重跑 guard。读取前核对身份和源哈希；旧观察不提供新的执行权限。 文件损坏或丢失不能自动重跑，选 Todo、lease 变化等仍需按原契约重新准入。 +The default ordinary quota CLI view may replace an exactly identical nested +`heartbeat_recommendation.blocked_priority_fallback` diagnostic with +`blocked_priority_fallback_ref="$.blocked_priority_fallback"`. Read the complete +root field in that same packet; resolving this reference needs no new guard or +detail command. Different mirrors remain inline. Raw decisions, saved captures +and the existing `--include-detail agent-todos` view retain their original shape. +Action dimensions remain equivalent; the raw source hash still identifies the +specific original rendering and must not be substituted across views. +The existing packet owner supplies this compatibility projection, keeping the +CLI from becoming another reader of the legacy guidance field. + +中文：普通 quota CLI 热路径将完全相同的 fallback 副本改为本包根字段引用，完整 +事实仍在当前包内，不需再次执行 guard。不同副本和既有详细视图保留；原始决定与 +capture 不变。动作维度等价不意味着不同渲染可以共用原始 source hash。 + ### Native Turn Host capture Executing `turn run-once --execute` now saves the original full decision before diff --git a/loopx/control_plane/quota/cli_projection.py b/loopx/control_plane/quota/cli_projection.py index 56fa998d7f..e1d50cf57e 100644 --- a/loopx/control_plane/quota/cli_projection.py +++ b/loopx/control_plane/quota/cli_projection.py @@ -3,6 +3,7 @@ from typing import Any from .decision_summary import compact_quota_decision +from .should_run_packet import reference_duplicate_fallback_guidance QUOTA_CLI_TODO_SUMMARY_COMPACTION_SCHEMA_VERSION = ( "quota_cli_todo_summary_compaction_v0" @@ -790,6 +791,7 @@ def compact_quota_should_run_cli_payload( compact = dict(compact) compact["goal_route_hint"] = _compact_goal_route_hint(goal_route_hint) compact = _compact_shadowed_action_projections(compact) + compact = reference_duplicate_fallback_guidance(compact) agent_lane_next_action = compact.get("agent_lane_next_action") if isinstance(agent_lane_next_action, dict) and "content_revision" in agent_lane_next_action: compact = dict(compact) diff --git a/loopx/control_plane/quota/should_run_packet.py b/loopx/control_plane/quota/should_run_packet.py index cec943ee86..e0e631becb 100644 --- a/loopx/control_plane/quota/should_run_packet.py +++ b/loopx/control_plane/quota/should_run_packet.py @@ -1,4 +1,5 @@ from __future__ import annotations +import json from .effective_action import EffectiveAction from ..effect_runtime import effect_runtime_result from .effect_program import ReceiptBoundReplayPhase @@ -152,6 +153,27 @@ ) +def reference_duplicate_fallback_guidance(payload: dict[str, Any]) -> dict[str, Any]: + """Reference a CLI-only mirror through the existing legacy packet owner.""" + fallback = payload.get("blocked_priority_fallback") + recommendation = payload.get("heartbeat_recommendation") + if not isinstance(fallback, dict) or not isinstance(recommendation, dict): + return payload + mirror = recommendation.get("blocked_priority_fallback") + if not isinstance(mirror, dict): + return payload + # Compare JSON types too: Python equality conflates False with 0. + if mirror is not fallback and ( + json.dumps(mirror, sort_keys=True, ensure_ascii=False) + != json.dumps(fallback, sort_keys=True, ensure_ascii=False) + ): + return payload + recommendation = dict(recommendation) + recommendation.pop("blocked_priority_fallback") + recommendation["blocked_priority_fallback_ref"] = "$.blocked_priority_fallback" + return {**payload, "heartbeat_recommendation": recommendation} + + def _interaction_runtime_root( runtime_root: str | Path | None, status_payload: Mapping[str, Any], diff --git a/tests/control_plane/test_quota_cli_projection.py b/tests/control_plane/test_quota_cli_projection.py index aa0d814e6e..7ae0e43846 100644 --- a/tests/control_plane/test_quota_cli_projection.py +++ b/tests/control_plane/test_quota_cli_projection.py @@ -17,6 +17,37 @@ ) +@pytest.mark.parametrize("mirror", ["shared", "identical", "different", "boolean-number", "missing-root"]) +def test_fallback_reference_preserves_complete_same_packet_facts(mirror: str) -> None: + fallback = {"notify_user": False, "blocked_items": [ + {"text": "保留完整原因和恢复条件。" * 200, "reason": "Wait for the original dependency"}, + ], "selected_executable": {"todo_id": "todo_fallback", "claimed_by": "worker"}} + nested = fallback if mirror == "shared" else deepcopy(fallback) + if mirror == "different": + nested["blocked_items"][0]["reason"] = "A different dependency" + elif mirror == "boolean-number": + nested["notify_user"] = 0 + payload = {"heartbeat_recommendation": {"notify": "DONT_NOTIFY", "agent_must_attempt": True, + "blocked_priority_fallback": nested}} + if mirror != "missing-root": + payload["blocked_priority_fallback"] = fallback + original = deepcopy(payload) + compact = compact_quota_should_run_cli_payload(payload) + recommendation = compact["heartbeat_recommendation"] + if mirror in {"shared", "identical"}: + assert "blocked_priority_fallback" not in recommendation + assert recommendation["blocked_priority_fallback_ref"] == "$.blocked_priority_fallback" + assert compact["blocked_priority_fallback"] == fallback + assert recommendation["agent_must_attempt"] is True + assert recommendation["notify"] == "DONT_NOTIFY" + assert len(json.dumps(compact, ensure_ascii=False)) < len(json.dumps(original, ensure_ascii=False)) + else: + assert recommendation["blocked_priority_fallback"] == nested + assert "blocked_priority_fallback_ref" not in recommendation + assert payload == original + assert compact_quota_should_run_cli_payload(payload, include_todo_summary_detail=True) == original + + @pytest.mark.parametrize("inline", ["absent", "identical", "different", "wrong-step"]) def test_replan_authoring_is_available_without_another_read(inline: str) -> None: authoring = {"schema_version": "goal_vision_replan_contract_v0", diff --git a/tests/control_plane/test_quota_decision_capture.py b/tests/control_plane/test_quota_decision_capture.py index 67c85c42f1..719db95122 100644 --- a/tests/control_plane/test_quota_decision_capture.py +++ b/tests/control_plane/test_quota_decision_capture.py @@ -67,6 +67,56 @@ def test_existing_symlink_is_not_followed(tmp_path): assert not (tmp_path / 'absent').exists() +@pytest.mark.parametrize('provider', ['file', 'sqlite']) +def test_real_cli_fallback_reference_keeps_capture_and_turn_effects(tmp_path, monkeypatch, provider): + from loopx.control_plane.coordination.runtime_shadow import build_todo_runtime_shadow_projection + from loopx.control_plane.quota.turn_envelope import quota_action_signature_document + from tests.control_plane.canonical_authority_fixture import ( + initialize_canonical_authority, isolate_sqlite_runtime, + ) + from tests.control_plane.test_quota_settlement_cli import _spend_run_count + + isolate_sqlite_runtime(tmp_path, monkeypatch) + project, runtime, registry = _write_fixture(tmp_path) + state = _configure_read_only_todo(project) + with state.open('a') as stream: + stream.write('\n- [ ] [P0] Preserve the original dependency and recovery evidence.\n' + f' \n') + code, listed = _run_cli(registry, runtime, 'todo', 'list', '--goal-id', GOAL_ID) + assert code == 0, listed + projection = build_todo_runtime_shadow_projection( + goal_id=GOAL_ID, handoff_mode='soft_claim', todos=listed['todos'], + ) + initialize_canonical_authority(runtime, GOAL_ID, projection, state_path=state, provider=provider) + args = ['quota', 'should-run', '--goal-id', GOAL_ID, '--agent-id', AGENT_ID, + '--todo-id', TODO_ID, '--turn-instance-id', TURN_ID, '--codex-app', + '--available-capability', 'shell', '--available-capability', 'filesystem_read'] + capture = tmp_path / 'original-decision' + code, ordinary = _run_cli(registry, runtime, *args, '--decision-output-dir', str(capture)) + assert code == 0, ordinary + saved = capture / 'decision.json' + original_bytes = saved.read_bytes() + full = json.loads(original_bytes) + fallback = full['blocked_priority_fallback'] + assert full['heartbeat_recommendation']['blocked_priority_fallback'] == fallback + assert ordinary['blocked_priority_fallback'] == fallback + assert ordinary['heartbeat_recommendation']['blocked_priority_fallback_ref'] == '$.blocked_priority_fallback' + assert 'blocked_priority_fallback' not in ordinary['heartbeat_recommendation'] + assert ordinary['heartbeat_receipt'] == full['heartbeat_receipt'] + assert ordinary['interaction_contract']['cli_channel']['settlement_plan'] == full[ + 'interaction_contract']['cli_channel']['settlement_plan'] + # Raw source hashes distinguish renderings; every action dimension must match. + assert quota_action_signature_document(ordinary) == quota_action_signature_document(full) + before = _heartbeat_receipt_count(runtime, TURN_ID) + code, replay = _run_cli(registry, runtime, *args) + assert code == 0, replay + assert _heartbeat_receipt_count(runtime, TURN_ID) == before + assert _spend_run_count(runtime) == 0 + assert saved.read_bytes() == original_bytes + + @pytest.mark.parametrize('provider', ['file', 'sqlite']) def test_real_cli_boundary_clauses_reach_shared_host_without_loss(tmp_path, monkeypatch, provider): from loopx.control_plane.coordination.runtime_shadow import build_todo_runtime_shadow_projection