@hunter-harness/workflow-harness 0.2.80 → 0.2.82

package/harness/bundles/general/codex/harness-test/SKILL.md

@@ -3,7 +3,7 @@ name: harness-test
 description: 测试执行：读取场景表，执行单元测试+API接口测试+数据兼容验证，输出测试报告。仅当用户显式调用 /harness-test
   时使用；不得在 run 结束后自动接续执行。
 ---
-<!-- generated by harness_deploy.py; core=b45619980b5b945d; overlay=none; agent=codex; do not edit -->
+<!-- generated by harness_deploy.py; core=2676178fa6f219fd; overlay=none; agent=codex; do not edit -->
 # harness-test — 测试执行
 
 ## Purpose
@@ -219,13 +219,23 @@ python <skills-root>/scripts/harness_test_guard.py record --project . --change-d
 
 `events.ndjson` 为唯一事实源（schema_version 3，兼容读取 v1/v2；`note` 承载人类可读摘要）；`logs/execution-log.md` 由 `harness_events.py` 渲染，**禁止用 Write/Edit 直接维护**。直接修改的内容会在 `phase.end` 或 finalize 时被完整重建覆盖，属于数据丢失；需要保留的详情必须进入事件 `note`。结构 → [[../protocols/report-pipeline-protocol.md|report-pipeline-protocol]]
 
+**`phase.start` 由 `harness_gate.py begin` 写，不要再手工追加一次。** 两条同 `run-id` 的
+`phase.start` 会让 `plan finalize` 以 `PHASE_START_DUPLICATE` 卡死，而且手工那次会先触发
+auto-seal、把正在开始的 attempt 封成 `RECOVERED`。要补触发指令说明就带 `--note` 跑 `gate begin`。
+（重复追加现已按 `(phase, run-id)` 判为幂等 no-op，但依赖它不如不写。）
+
 ```powershell
-python <skills-root>/scripts/harness_events.py append --change-dir ".harness/changes/<change-name>" --phase <phase> --type phase.start --run-id <phase-run-id> --attempt <n> --note "<触发指令>"
+# 阶段开始：gate begin 负责，note 在这里给
+python <skills-root>/scripts/harness_gate.py begin --change-dir ".harness/changes/<change-name>" --phase <phase> --note "<触发指令>"
+# 阶段中的其他事件才用 append
+python <skills-root>/scripts/harness_events.py append --change-dir ".harness/changes/<change-name>" --phase <phase> --type <command|issue|verification> --run-id <phase-run-id> --note "<摘要>"
 ```
 
 > **脚本接线**：`harness_events.py append`；`harness_archive.py finalize`；`harness_preflight.py check`；`harness_ledger.py can-reuse`；`harness_service.py ensure/stop`（须 `--files`/`serviceStart.inputFiles`）。JSON 输出按 D13 护栏解读。
 
 > **Task 4 §6.1 写入契约**：普通 `append` = 加锁 -> 追加一行 -> fsync -> 解锁，**不 load 历史、不渲染**（O(1)，跨进程锁 `events.ndjson.lock`，UUID 用完整 `uuid4().hex` 无需去重扫描）。仅 `--type phase.end` append 在追加成功后渲染一次 `execution-log.md`；显式 `harness_events.py render` 随时从完整 events 重建；`harness_archive.py finalize` 在 collect 前强制 render 一次。高频 command append 期间 log 可能滞后，phase 边界保持最新。
 
-每个阶段的 `phase.start` 与对应 `phase.end` 必须复用同一 `--run-id` / `--attempt`；阶段结束必须写 `--status OK|WARN|FAIL|BLOCKED`。重试同一阶段时生成新的 run-id 并增加 `--attempt <n>`，不得覆盖或伪装成一次执行。跨工具继续执行时写 `--executor-tool <codex|claude-code|codebuddy|cursor>`，并在接棒事件写 `--handoff-from-tool` / `--handoff-reason`；也可由 `HUNTER_HARNESS_TOOL/AGENT/MODEL/RUN_ID` 环境变量统一注入。
+每个阶段的 `phase.start` 与对应 `phase.end` 必须复用同一 `--run-id` / `--attempt`；阶段结束必须写 `--status OK|WARN|FAIL|BLOCKED`。重试同一阶段时生成新的 run-id 并增加 `--attempt <n>`，不得覆盖或伪装成一次执行。**`attempt` 按 phase 全局递增，不是按 run-id**：一个 run-id 只绑定一个 attempt，重试必须「新 run-id ＋ 下一个 attempt」两者同时换，只换其一会撞 `EVENT_ATTEMPT_CONFLICT` 或 `PHASE_ALREADY_CLOSED`。已发布的 plan 用 `harness_plan_finalize.py republish` 自动分配两者。
+
+阶段跑得久（plan/run 常见）时用 `harness_context.py renew --project . --change <cn> --executor <tool>` 续租；租约到期本身不再阻断 `close`（同一 owner 的过期租约不构成冲突，收据里记 `leaseLapsed`），但续租能让 `view` 的状态如实反映在跑。跨工具继续执行时写 `--executor-tool <codex|claude-code|codebuddy|cursor>`，并在接棒事件写 `--handoff-from-tool` / `--handoff-reason`；也可由 `HUNTER_HARNESS_TOOL/AGENT/MODEL/RUN_ID` 环境变量统一注入。
 > 片段：logging · phase=`test` · 事件：phase/command/verification/decision/issue/artifact

package/harness/bundles/general/codex/scripts/harness_archive.py

@@ -9041,7 +9041,27 @@ def collect_archive_core_paths(project_root: Path, archive_dir: Path) -> list[st
     return [source.relative_to(root).as_posix() for source, *_ in file_specs]
 
 
-def _archive_source_identity(project_root: Path) -> dict[str, str | None]:
+def _archive_recorded_commit(archive_dir: Path | None) -> str | None:
+    """The commit the archive itself recorded as its endpoint, if any."""
+    if archive_dir is None:
+        return None
+    try:
+        summary = read_json(archive_dir / "reports" / "final" / "summary-data.json")
+    except (OSError, ValueError, json.JSONDecodeError):
+        return None
+    if not isinstance(summary, dict):
+        return None
+    for key in ("finalCommit", "final_commit", "currentHead", "headCommit"):
+        value = summary.get(key)
+        if isinstance(value, str) and re.fullmatch(r"[0-9a-f]{40}", value.strip()):
+            return value.strip()
+    return None
+
+
+def _archive_source_identity(
+    project_root: Path,
+    archive_dir: Path | None = None,
+) -> dict[str, str | None]:
     def git_value(*args: str) -> str | None:
         try:
             completed = subprocess.run(
@@ -9060,6 +9080,15 @@ def _archive_source_identity(project_root: Path) -> dict[str, str | None]:
         value = completed.stdout.strip()
         return value if completed.returncode == 0 and value else None
 
+    # Prefer the archive's own endpoint over live HEAD so the package stays a
+    # pure function of the sealed directory. Falling back to HEAD keeps older
+    # archives that never recorded a commit working exactly as before.
+    recorded = _archive_recorded_commit(archive_dir)
+    if recorded is not None:
+        return {
+            "commit": recorded,
+            "tree": git_value("rev-parse", f"{recorded}^{{tree}}"),
+        }
     return {
         "commit": git_value("rev-parse", "HEAD"),
         "tree": git_value("rev-parse", "HEAD^{tree}"),
@@ -9244,7 +9273,7 @@ def build_archive_package(
         "profile": "core-v1",
         "change_key": change_key,
         "created_at": _archive_created_at(summary_path),
-        "source": _archive_source_identity(project_root),
+        "source": _archive_source_identity(project_root, archive_dir),
         "files": entries,
     }
     manifest_bytes = (
@@ -9482,35 +9511,83 @@ def auto_push_managed_snapshot(project_root: Path) -> dict[str, Any]:
     }
 
 
+def load_retained_package(project_root: Path, change_key: str) -> dict[str, Any] | None:
+    """A package kept on disk for retry, described by its own pending receipt.
+
+    Retrying is only meaningful against the bytes the server already saw. A
+    rebuild is a different package: the archive directory and the harness both
+    move on, so "retry the same ZIP" has to mean exactly that.
+    """
+    packages = project_root / ".harness" / "state" / "local" / "archive-packages"
+    zip_path = packages / f"{change_key}.zip"
+    receipt_path = packages / f"{change_key}.upload.json"
+    if not zip_path.is_file() or not receipt_path.is_file():
+        return None
+    try:
+        receipt = read_json(receipt_path)
+    except (OSError, ValueError, json.JSONDecodeError):
+        return None
+    if not isinstance(receipt, dict):
+        return None
+    recorded = str(receipt.get("packageSha256") or "")
+    actual = "sha256:" + hashlib.sha256(zip_path.read_bytes()).hexdigest()
+    if recorded and recorded != actual:
+        # The receipt no longer describes the file beside it; refuse to guess.
+        return None
+    return {
+        "schemaVersion": 1,
+        "profile": "core-v1",
+        "changeKey": change_key,
+        "packagePath": str(zip_path),
+        "packageSha256": actual,
+        "manifestSha256": receipt.get("manifestSha256"),
+        "fileCount": receipt.get("fileCount"),
+        "sizeBytes": zip_path.stat().st_size,
+        "paths": list(receipt.get("paths") or []),
+        "knowledgeCandidateCount": receipt.get("knowledgeCandidateCount"),
+        "reused": True,
+    }
+
+
 def auto_push_archive_core(
     project_root: Path,
     archive_dir: Path,
     *,
     change_key: str | None = None,
     extra_entries: dict[str, tuple[str, str, bytes]] | None = None,
+    reuse_retained_package: bool = False,
 ) -> dict[str, Any]:
     """Build and upload one deterministic core ZIP after finalize.
 
     The ZIP and its per-change receipt are always created before auth is
     evaluated. Failures become warnings so archive success is never rolled
     back by missing credentials or a remote outage.
+
+    With ``reuse_retained_package`` a package already kept for retry is uploaded
+    as-is instead of being rebuilt, which is what the retry messages promise.
     """
     project_root = project_root.resolve()
     effective_change_key = change_key or archive_dir.name
-    try:
-        package = build_archive_package(
-            project_root,
-            archive_dir,
-            effective_change_key,
-            extra_entries=extra_entries,
-        )
-    except (OSError, ValueError, UnicodeDecodeError, json.JSONDecodeError) as exc:
-        return {
-            "skipped": False,
-            "ok": False,
-            "reasonCode": "ARCHIVE_PACKAGE_BUILD_FAILED",
-            "warning": f"核心归档 ZIP 生成失败：{exc}",
-        }
+    package = (
+        load_retained_package(project_root, effective_change_key)
+        if reuse_retained_package
+        else None
+    )
+    if package is None:
+        try:
+            package = build_archive_package(
+                project_root,
+                archive_dir,
+                effective_change_key,
+                extra_entries=extra_entries,
+            )
+        except (OSError, ValueError, UnicodeDecodeError, json.JSONDecodeError) as exc:
+            return {
+                "skipped": False,
+                "ok": False,
+                "reasonCode": "ARCHIVE_PACKAGE_BUILD_FAILED",
+                "warning": f"核心归档 ZIP 生成失败：{exc}",
+            }
     core_paths = list(package["paths"])
     package_path = Path(str(package["packagePath"]))
     try:
@@ -10346,6 +10423,7 @@ def cmd_republish(
     project_root: Path | None,
     dry_run: bool,
     inject_knowledge: bool = True,
+    retry_retained: bool = False,
 ) -> tuple[int, dict[str, Any]]:
     """Rebuild and re-upload the core ZIP for an already sealed archive.
 
@@ -10405,6 +10483,70 @@ def cmd_republish(
             extra_entries = {}
     payload["knowledgeCandidatesInjected"] = "candidates/knowledge.json" in extra_entries
 
+    if retry_retained:
+        retained = load_retained_package(root, change_key)
+        if retained is None:
+            payload.update({
+                "ok": False,
+                "reasonCode": "ARCHIVE_RETAINED_PACKAGE_UNAVAILABLE",
+                "error": (
+                    "没有可重试的留存包：需要 "
+                    f".harness/state/local/archive-packages/{change_key}.zip 与同名 "
+                    ".upload.json 同时存在，且回执记录的 packageSha256 与 ZIP 字节一致。"
+                ),
+                "nextAction": "去掉 --retry-retained 可按封存目录重建一个新包。",
+            })
+            return 1, payload
+        payload["retainedPackage"] = {
+            "packagePath": retained["packagePath"],
+            "packageSha256": retained["packageSha256"],
+            "fileCount": retained["fileCount"],
+            "sizeBytes": retained["sizeBytes"],
+        }
+        durable = read_durable_archive_receipt(root, change_key)
+        stored_sha = str((durable or {}).get("packageSha256") or "")
+        if (
+            durable is not None
+            and str(durable.get("archiveStatus") or "") == "durable"
+            and stored_sha
+            and stored_sha != retained["packageSha256"]
+        ):
+            # Identical bytes stay allowed below — re-uploading the same package
+            # is the documented way to retry a failed knowledge index. Different
+            # bytes are refused by the server, so do not spend the upload.
+            payload.update({
+                "ok": False,
+                "reasonCode": "ARCHIVE_REMOTE_IMMUTABLE_CONFLICT",
+                "remoteReceipt": {
+                    "archiveId": durable.get("archiveId"),
+                    "packageSha256": stored_sha,
+                },
+                "error": (
+                    "留存包与远端已发布的包字节不同，服务端对同一 change key 只保存一个"
+                    "不可变包，不接受替换。"
+                ),
+                "nextAction": (
+                    "该留存包是一次失败尝试的残留，不是已发布的那一份；"
+                    "确认无用后可删除 .zip 与 .upload.json。"
+                ),
+            })
+            return 1, payload
+        if dry_run:
+            payload.update({"ok": True, "reasonCode": "ARCHIVE_RETAINED_PACKAGE_PREVIEW"})
+            return 0, payload
+        result = auto_push_archive_core(
+            root,
+            archive_dir,
+            change_key=change_key,
+            reuse_retained_package=True,
+        )
+        payload["push"] = result
+        payload["ok"] = bool(result.get("ok"))
+        payload["reasonCode"] = result.get("reasonCode")
+        if result.get("warning"):
+            payload.setdefault("warnings", []).append(str(result["warning"]))
+        return 0 if payload["ok"] else 1, payload
+
     # Build once, into a preview path, so the bytes can be compared against any
     # durable receipt before an upload is attempted. The builder refuses to write
     # outside the project, so the preview lands beside the real packages under a
@@ -10479,8 +10621,8 @@ def cmd_republish(
                     f"{package['knowledgeCandidateCount']} 条知识候选"
                     "（原包上传时还没有 candidates/knowledge.json），因此字节必然不同——"
                     "已发布归档无法从客户端追加知识条目，需要平台侧提供重新索引或归档版本化能力。"
-                    "若只是想重试同一个包（例如知识索引失败），加 --no-knowledge-injection "
-                    "可重建与封存目录完全一致的字节。"
+                    "若只是想重试同一个包（例如知识索引失败），用 --retry-retained "
+                    "上传盘上留存的原包字节；重建得不到已发布的字节。"
                     if payload.get("knowledgeCandidatesInjected")
                     else "本地归档内容与已上传的包不一致，请先确认哪一份才是应保留的事实。"
                 )
@@ -10513,6 +10655,7 @@ def cmd_republish_cli(args: argparse.Namespace) -> int:
         project_root=resolve_path(args.project) if args.project else None,
         dry_run=bool(args.dry_run),
         inject_knowledge=not bool(args.no_knowledge_injection),
+        retry_retained=bool(args.retry_retained),
     )
     emit_json(payload)
     return code
@@ -10696,12 +10839,22 @@ def build_parser() -> argparse.ArgumentParser:
         action="store_true",
         help="build the package and report its contents without uploading",
     )
+    p_republish.add_argument(
+        "--retry-retained",
+        action="store_true",
+        help=(
+            "upload the package already kept on disk for retry, byte for byte, "
+            "instead of rebuilding it"
+        ),
+    )
     p_republish.add_argument(
         "--no-knowledge-injection",
         action="store_true",
         help=(
-            "rebuild exactly what the sealed archive holds; needed to retry the "
-            "same bytes the server already stored"
+            "rebuild without injecting knowledge candidates. Note this does NOT "
+            "reproduce an already-published package: the manifest also binds the "
+            "archive's own commit and the sealed directory may have moved on. To "
+            "retry the exact bytes the server saw, use --retry-retained"
         ),
     )
     p_republish.add_argument("--json", action="store_true", default=True)

package/harness/bundles/general/codex/scripts/harness_context.py

@@ -1165,9 +1165,11 @@ def close_transition(
             "holder": lease.get("owner"),
             "leasePhase": lease.get("phase"),
         }
+    # Owner and phase already matched above, so nobody took this lease over: a
+    # takeover rewrites the file and would have produced CONTEXT_LEASE_MISMATCH.
+    # A lapsed deadline therefore proves only that the phase outlived the TTL.
     expiry = _parse_time(lease.get("expiresAt"))
-    if expiry is None or _now() >= expiry:
-        return {"ok": False, "code": "CONTEXT_LEASE_EXPIRED"}
+    lease_lapsed = expiry is None or _now() >= expiry
 
     previous_hash = transitions[-1].get("receiptHash") if transitions else None
     attempt = (
@@ -1192,6 +1194,12 @@ def close_transition(
         "closedAt": _now().isoformat(),
         "previousReceiptHash": previous_hash,
     }
+    if lease_lapsed:
+        # Recorded, not hidden: the phase ran past its lease without a renewal.
+        receipt["leaseLapsed"] = {
+            "expiresAt": lease.get("expiresAt"),
+            "acquiredAt": lease.get("acquiredAt"),
+        }
     receipt["receiptHash"] = _payload_hash(receipt)
     _append_ndjson(paths["transitions"], receipt)
     paths["lease"].unlink(missing_ok=True)
@@ -1267,7 +1275,11 @@ def _begin_transition_unlocked(
             "legacyBootstrap": True,
             "message": (
                 "没有交接凭证。v2 计划应有 committed 的 meta/publication-journals/*.json "
-                "作为补录证据；legacy 计划需先完成 plan 阶段的 close"
+                "作为补录证据；legacy 计划需先完成 plan 阶段的 close。"
+                "首次进入某阶段（还没有任何 transition 凭证）时不要用 context begin——"
+                "改跑 `harness_gate.py begin --change-dir .harness/changes/<change> "
+                "--phase <phase> --run-id <phase>_<uuid4hex> --note \"<触发指令>\"`，"
+                "它会领取租约并写好 phase.start。"
             ),
         }
     receipt = transitions[-1]
@@ -1458,6 +1470,53 @@ def cancel_prepared_context(
     }
 
 
+def renew_lease(
+    project: Path,
+    change: str,
+    *,
+    executor: str,
+    ttl_seconds: int = 3600,
+) -> dict[str, Any]:
+    """Extend the current phase lease in place.
+
+    Only the recorded owner may renew, and only while no other executor has
+    taken the lease over. Renewal touches nothing but ``expiresAt``, so it is
+    safe to call on a heartbeat during a long phase.
+    """
+    project = project.resolve()
+    try:
+        _, _, state_root = _contract(project, change)
+    except ValueError as exc:
+        return {"ok": False, "code": _contract_error_code(exc), "error": str(exc)}
+    paths = _paths(state_root)
+    if not paths["lease"].is_file():
+        return {"ok": False, "code": "CONTEXT_LEASE_REQUIRED"}
+    with _exclusive_state_lock(paths["runtime"] / "branch-selection.lock"):
+        try:
+            lease = _read_json(paths["lease"])
+        except (OSError, ValueError, json.JSONDecodeError):
+            return {"ok": False, "code": "CONTEXT_LEASE_INVALID"}
+        if lease.get("owner") != executor:
+            return {
+                "ok": False,
+                "code": "CONTEXT_LEASE_MISMATCH",
+                "holder": lease.get("owner"),
+            }
+        expiry = _parse_time(lease.get("expiresAt"))
+        lapsed = expiry is None or _now() >= expiry
+        lease["expiresAt"] = (
+            _now() + dt.timedelta(seconds=max(1, int(ttl_seconds)))
+        ).isoformat()
+        lease["renewedAt"] = _now().isoformat()
+        _write_json_atomic(paths["lease"], lease)
+    return {
+        "ok": True,
+        "code": "CONTEXT_LEASE_RENEWED",
+        "lease": lease,
+        "wasLapsed": lapsed,
+    }
+
+
 def context_view(project: Path, change: str) -> dict[str, Any]:
     project = Path(project).resolve()
     try:
@@ -1536,6 +1595,14 @@ def build_parser() -> argparse.ArgumentParser:
     begin.add_argument("--change", required=True)
     begin.add_argument("--phase", required=True)
     begin.add_argument("--executor", required=True)
+    renew = sub.add_parser(
+        "renew", help="extend the current phase lease without touching context"
+    )
+    renew.add_argument("--json", action="store_true")
+    renew.add_argument("--project", required=True, type=Path)
+    renew.add_argument("--change", required=True)
+    renew.add_argument("--executor", required=True)
+    renew.add_argument("--ttl-seconds", type=int, default=3600)
     configure = sub.add_parser("configure-plan")
     configure.add_argument("--json", action="store_true")
     configure.add_argument("--project", required=True, type=Path)
@@ -1588,6 +1655,13 @@ def main(argv: list[str] | None = None) -> int:
             phase=args.phase,
             executor=args.executor,
         )
+    elif args.command == "renew":
+        result = renew_lease(
+            args.project,
+            args.change,
+            executor=args.executor,
+            ttl_seconds=args.ttl_seconds,
+        )
     elif args.command == "configure-plan":
         result = configure_phase_plan(
             args.project,

package/harness/bundles/general/codex/scripts/harness_events.py

@@ -1904,6 +1904,7 @@ def cmd_append(args: argparse.Namespace) -> int:
             file=sys.stderr,
         )
     phase_closed = False
+    phase_start_duplicate = False
     projection_error: str | None = None
     attempt_error: str | None = None
     auto_sealed_events: list[dict[str, Any]] = []
@@ -1951,10 +1952,25 @@ def cmd_append(args: argparse.Namespace) -> int:
                 except ValueError as exc:
                     projection_error = str(exc)
             elif args.type == "phase.start":
+                # Re-appending the same (phase, run_id) is the caller restating a
+                # start that already exists — `gate begin` writes one, and the
+                # logging protocol used to ask for a second. Treat it as a no-op
+                # rather than a duplicate; two identical phase.start events make
+                # `plan finalize` unrecoverable via PHASE_START_DUPLICATE.
+                if run_id is not None and any(
+                    item.get("phase") == args.phase
+                    and item.get("type") == "phase.start"
+                    and item.get("run_id") == run_id
+                    for item in existing_events
+                ):
+                    phase_start_duplicate = True
                 # HH-WF-20260730-001: a new phase.start must not be appended
                 # while a prior attempt for the same phase is still open —
                 # auto-seal it first so timing never has two open attempts.
-                open_attempts = open_attempts_for_phase(existing_events, args.phase)
+                open_attempts = (
+                    [] if phase_start_duplicate
+                    else open_attempts_for_phase(existing_events, args.phase)
+                )
                 if open_attempts:
                     inferred_reason = infer_auto_seal_reason(
                         open_attempts[-1].get("events") or []
@@ -1964,7 +1980,12 @@ def cmd_append(args: argparse.Namespace) -> int:
                         phase=args.phase,
                         seal_reason=inferred_reason,
                     )
-            if not phase_closed and projection_error is None and attempt_error is None:
+            if (
+                not phase_closed
+                and not phase_start_duplicate
+                and projection_error is None
+                and attempt_error is None
+            ):
                 for seal_event in auto_sealed_events:
                     atomic_append_line(
                         path,
@@ -1976,6 +1997,18 @@ def cmd_append(args: argparse.Namespace) -> int:
                 )
     except (OSError, TimeoutError, ValueError) as exc:
         return emit_error(f"append failed: {exc}", as_json=as_json)
+    if phase_start_duplicate:
+        if as_json:
+            print(json.dumps({
+                "ok": True,
+                "skipped": True,
+                "reason": "phase-start-already-recorded",
+                "phase": args.phase,
+                "run_id": event.get("run_id"),
+            }, ensure_ascii=False))
+        else:
+            print("ok (phase.start already recorded for this run id)")
+        return 0
     if phase_closed:
         return emit_error(
             "PHASE_ALREADY_CLOSED: refusing a second phase.end for the same "

package/harness/bundles/general/codex/scripts/harness_plan_finalize.py

@@ -735,7 +735,12 @@ def _validate_plan_start(
     if len(matching) > 1:
         return _result_error(
             "PHASE_START_DUPLICATE",
-            f"found {len(matching)} matching plan phase.start events",
+            f"found {len(matching)} matching plan phase.start events for run id "
+            f"{run_id}. `harness_gate.py begin` already appends phase.start; a "
+            "second manual append of the same run id creates this state. Recover "
+            "with a fresh run id and the next attempt number (attempts are global "
+            "per phase, not per run id): append phase.start with both, then "
+            "finalize with the same pair.",
         )
     return {"ok": True, "phaseStartCount": 1}
 

package/harness/bundles/general/cursor/.harness-build.json

@@ -2,5 +2,5 @@
   "schemaVersion": 1,
   "overlay": "none",
   "agent": "cursor",
-  "coreHash": "629c995f68a2e765"
+  "coreHash": "bd7ebe189b5b88e3"
 }

package/harness/bundles/general/cursor/harness-archive/SKILL.md

@@ -3,7 +3,7 @@ name: harness-archive
 description: 归档所有变更产出（计划/测试报告/审查/SQL/API文档）到 .harness/archive/，含归档元数据。仅当用户显式调用
   /harness-archive 时使用；不得被其他阶段 skill 自动接续触发。
 ---
-<!-- generated by harness_deploy.py; core=629c995f68a2e765; overlay=none; agent=cursor; do not edit -->
+<!-- generated by harness_deploy.py; core=bd7ebe189b5b88e3; overlay=none; agent=cursor; do not edit -->
 # harness-archive — 变更归档
 
 ## Purpose
@@ -157,7 +157,11 @@ python <skills-root>/scripts/harness_archive.py republish --change <change-key>
 （exit 0，不重传）；字节不同 → `ARCHIVE_REMOTE_IMMUTABLE_CONFLICT`（exit 1，**不发起上传**）。
 **已发布归档无法从客户端补知识条目**——注入候选必然改变字节，服务端拒绝替换；要补需要平台侧
 提供重新索引或归档版本化能力。
-只想重试**同一个包**（例如知识索引失败）时加 `--no-knowledge-injection`，它按封存目录原样重建，字节与已存包一致。
+只想重试**同一个包**（例如知识索引失败）时用 `--retry-retained`：它上传盘上留存的原包字节，
+不重建。**重建得不到已发布的字节**——包 manifest 绑定归档自己的提交，封存目录与 harness 本身
+也都会前进；`--no-knowledge-injection` 只是不注入候选，不等于能复现旧包。
+留存包与远端字节不同（即那是一次失败尝试的残留）时，`--retry-retained` 同样在本地判出
+`ARCHIVE_REMOTE_IMMUTABLE_CONFLICT` 并拒绝上传。
 
 ### 三、文件移动只用内置工具或 PowerShell
 
@@ -258,13 +262,23 @@ git 命令通过 `powershell.exe -Command "..."` 执行；archive-meta.md 和 su
 
 `events.ndjson` 为唯一事实源（schema_version 3，兼容读取 v1/v2；`note` 承载人类可读摘要）；`logs/execution-log.md` 由 `harness_events.py` 渲染，**禁止用 Write/Edit 直接维护**。直接修改的内容会在 `phase.end` 或 finalize 时被完整重建覆盖，属于数据丢失；需要保留的详情必须进入事件 `note`。结构 → [[../protocols/report-pipeline-protocol.md|report-pipeline-protocol]]
 
+**`phase.start` 由 `harness_gate.py begin` 写，不要再手工追加一次。** 两条同 `run-id` 的
+`phase.start` 会让 `plan finalize` 以 `PHASE_START_DUPLICATE` 卡死，而且手工那次会先触发
+auto-seal、把正在开始的 attempt 封成 `RECOVERED`。要补触发指令说明就带 `--note` 跑 `gate begin`。
+（重复追加现已按 `(phase, run-id)` 判为幂等 no-op，但依赖它不如不写。）
+
 ```powershell
-python <skills-root>/scripts/harness_events.py append --change-dir ".harness/changes/<change-name>" --phase <phase> --type phase.start --run-id <phase-run-id> --attempt <n> --note "<触发指令>"
+# 阶段开始：gate begin 负责，note 在这里给
+python <skills-root>/scripts/harness_gate.py begin --change-dir ".harness/changes/<change-name>" --phase <phase> --note "<触发指令>"
+# 阶段中的其他事件才用 append
+python <skills-root>/scripts/harness_events.py append --change-dir ".harness/changes/<change-name>" --phase <phase> --type <command|issue|verification> --run-id <phase-run-id> --note "<摘要>"
 ```
 
 > **脚本接线**：`harness_events.py append`；`harness_archive.py finalize`；`harness_preflight.py check`；`harness_ledger.py can-reuse`；`harness_service.py ensure/stop`（须 `--files`/`serviceStart.inputFiles`）。JSON 输出按 D13 护栏解读。
 
 > **Task 4 §6.1 写入契约**：普通 `append` = 加锁 -> 追加一行 -> fsync -> 解锁，**不 load 历史、不渲染**（O(1)，跨进程锁 `events.ndjson.lock`，UUID 用完整 `uuid4().hex` 无需去重扫描）。仅 `--type phase.end` append 在追加成功后渲染一次 `execution-log.md`；显式 `harness_events.py render` 随时从完整 events 重建；`harness_archive.py finalize` 在 collect 前强制 render 一次。高频 command append 期间 log 可能滞后，phase 边界保持最新。
 
-每个阶段的 `phase.start` 与对应 `phase.end` 必须复用同一 `--run-id` / `--attempt`；阶段结束必须写 `--status OK|WARN|FAIL|BLOCKED`。重试同一阶段时生成新的 run-id 并增加 `--attempt <n>`，不得覆盖或伪装成一次执行。跨工具继续执行时写 `--executor-tool <codex|claude-code|codebuddy|cursor>`，并在接棒事件写 `--handoff-from-tool` / `--handoff-reason`；也可由 `HUNTER_HARNESS_TOOL/AGENT/MODEL/RUN_ID` 环境变量统一注入。
+每个阶段的 `phase.start` 与对应 `phase.end` 必须复用同一 `--run-id` / `--attempt`；阶段结束必须写 `--status OK|WARN|FAIL|BLOCKED`。重试同一阶段时生成新的 run-id 并增加 `--attempt <n>`，不得覆盖或伪装成一次执行。**`attempt` 按 phase 全局递增，不是按 run-id**：一个 run-id 只绑定一个 attempt，重试必须「新 run-id ＋ 下一个 attempt」两者同时换，只换其一会撞 `EVENT_ATTEMPT_CONFLICT` 或 `PHASE_ALREADY_CLOSED`。已发布的 plan 用 `harness_plan_finalize.py republish` 自动分配两者。
+
+阶段跑得久（plan/run 常见）时用 `harness_context.py renew --project . --change <cn> --executor <tool>` 续租；租约到期本身不再阻断 `close`（同一 owner 的过期租约不构成冲突，收据里记 `leaseLapsed`），但续租能让 `view` 的状态如实反映在跑。跨工具继续执行时写 `--executor-tool <codex|claude-code|codebuddy|cursor>`，并在接棒事件写 `--handoff-from-tool` / `--handoff-reason`；也可由 `HUNTER_HARNESS_TOOL/AGENT/MODEL/RUN_ID` 环境变量统一注入。
 > 片段：logging · phase=`archive`

package/harness/bundles/general/cursor/harness-codebase-map/SKILL.md

@@ -3,7 +3,7 @@ name: harness-codebase-map
 description: 分析现有代码库并生成 .harness/codebase/map/
   下的结构化代码库地图。用于陌生项目理解、重大改造前梳理、harness-sync 提示 map 过期后的人工刷新。
 ---
-<!-- generated by harness_deploy.py; core=629c995f68a2e765; overlay=none; agent=cursor; do not edit -->
+<!-- generated by harness_deploy.py; core=bd7ebe189b5b88e3; overlay=none; agent=cursor; do not edit -->
 # harness-codebase-map — 代码库地图
 
 ## Purpose
@@ -94,13 +94,23 @@ description: 分析现有代码库并生成 .harness/codebase/map/
 
 `events.ndjson` 为唯一事实源（schema_version 3，兼容读取 v1/v2；`note` 承载人类可读摘要）；`logs/execution-log.md` 由 `harness_events.py` 渲染，**禁止用 Write/Edit 直接维护**。直接修改的内容会在 `phase.end` 或 finalize 时被完整重建覆盖，属于数据丢失；需要保留的详情必须进入事件 `note`。结构 → [[../protocols/report-pipeline-protocol.md|report-pipeline-protocol]]
 
+**`phase.start` 由 `harness_gate.py begin` 写，不要再手工追加一次。** 两条同 `run-id` 的
+`phase.start` 会让 `plan finalize` 以 `PHASE_START_DUPLICATE` 卡死，而且手工那次会先触发
+auto-seal、把正在开始的 attempt 封成 `RECOVERED`。要补触发指令说明就带 `--note` 跑 `gate begin`。
+（重复追加现已按 `(phase, run-id)` 判为幂等 no-op，但依赖它不如不写。）
+
 ```powershell
-python <skills-root>/scripts/harness_events.py append --change-dir ".harness/changes/<change-name>" --phase <phase> --type phase.start --run-id <phase-run-id> --attempt <n> --note "<触发指令>"
+# 阶段开始：gate begin 负责，note 在这里给
+python <skills-root>/scripts/harness_gate.py begin --change-dir ".harness/changes/<change-name>" --phase <phase> --note "<触发指令>"
+# 阶段中的其他事件才用 append
+python <skills-root>/scripts/harness_events.py append --change-dir ".harness/changes/<change-name>" --phase <phase> --type <command|issue|verification> --run-id <phase-run-id> --note "<摘要>"
 ```
 
 > **脚本接线**：`harness_events.py append`；`harness_archive.py finalize`；`harness_preflight.py check`；`harness_ledger.py can-reuse`；`harness_service.py ensure/stop`（须 `--files`/`serviceStart.inputFiles`）。JSON 输出按 D13 护栏解读。
 
 > **Task 4 §6.1 写入契约**：普通 `append` = 加锁 -> 追加一行 -> fsync -> 解锁，**不 load 历史、不渲染**（O(1)，跨进程锁 `events.ndjson.lock`，UUID 用完整 `uuid4().hex` 无需去重扫描）。仅 `--type phase.end` append 在追加成功后渲染一次 `execution-log.md`；显式 `harness_events.py render` 随时从完整 events 重建；`harness_archive.py finalize` 在 collect 前强制 render 一次。高频 command append 期间 log 可能滞后，phase 边界保持最新。
 
-每个阶段的 `phase.start` 与对应 `phase.end` 必须复用同一 `--run-id` / `--attempt`；阶段结束必须写 `--status OK|WARN|FAIL|BLOCKED`。重试同一阶段时生成新的 run-id 并增加 `--attempt <n>`，不得覆盖或伪装成一次执行。跨工具继续执行时写 `--executor-tool <codex|claude-code|codebuddy|cursor>`，并在接棒事件写 `--handoff-from-tool` / `--handoff-reason`；也可由 `HUNTER_HARNESS_TOOL/AGENT/MODEL/RUN_ID` 环境变量统一注入。
+每个阶段的 `phase.start` 与对应 `phase.end` 必须复用同一 `--run-id` / `--attempt`；阶段结束必须写 `--status OK|WARN|FAIL|BLOCKED`。重试同一阶段时生成新的 run-id 并增加 `--attempt <n>`，不得覆盖或伪装成一次执行。**`attempt` 按 phase 全局递增，不是按 run-id**：一个 run-id 只绑定一个 attempt，重试必须「新 run-id ＋ 下一个 attempt」两者同时换，只换其一会撞 `EVENT_ATTEMPT_CONFLICT` 或 `PHASE_ALREADY_CLOSED`。已发布的 plan 用 `harness_plan_finalize.py republish` 自动分配两者。
+
+阶段跑得久（plan/run 常见）时用 `harness_context.py renew --project . --change <cn> --executor <tool>` 续租；租约到期本身不再阻断 `close`（同一 owner 的过期租约不构成冲突，收据里记 `leaseLapsed`），但续租能让 `view` 的状态如实反映在跑。跨工具继续执行时写 `--executor-tool <codex|claude-code|codebuddy|cursor>`，并在接棒事件写 `--handoff-from-tool` / `--handoff-reason`；也可由 `HUNTER_HARNESS_TOOL/AGENT/MODEL/RUN_ID` 环境变量统一注入。
 > 片段：logging · phase=`codebase-map`；默认写 reports，有未归档变更时另写 change events