Allow skill synchronization without session uploads

README.md

@@ -515,6 +515,8 @@ To verify the loop:
 2. Use the agent through its local proxy. After session finalization (the default idle timeout is 180 seconds), confirm a session artifact reaches `team-lan/sessions/` in the share. The evolver may consume it on the next cycle.
 3. Inspect server logs and `evolve_history.jsonl` after a processing cycle. Evolution depends on usable evidence and is not guaranteed to create a new skill every cycle. Accepted changes appear under `team-lan/skills/`; each client's poll then pulls them locally.
 
+To synchronize skills without exporting conversation/tool contents, run `skillclaw config sharing.upload_sessions false` and restart the proxy. This disables both periodic and final session uploads, including uploads to a local shared directory. Skill push/pull remains available. Existing shared artifacts, local records, and requests to your inference provider are unaffected. New setup asks explicitly (default: no); existing configurations keep their previous upload behavior until changed.
+
 This setup shares session evidence and generated skills with the group. `group_id` is a namespace, not an access-control boundary; use filesystem/share permissions and separate roots for different trust groups. Automated privacy redaction is not provided by this setup. The default publish mode is `direct`; use the `validated` flow above if clients should review candidates before publication.
 
 For Cloudflare R2, use the **account endpoint**, such as `https://<ACCOUNT_ID>.r2.cloudflarestorage.com`, with `sharing.backend s3`, `sharing.bucket <bucket-name>`, and `sharing.region auto`. Do not append the bucket to `sharing.endpoint`: boto3 sends the bucket separately. SkillClaw rejects that configuration before making a storage request. See the [R2 S3 API documentation](https://developers.cloudflare.com/r2/api/s3/api/).

assets/README_ZH.md

@@ -465,6 +465,8 @@ Bob 使用相同配置，只把 `sharing.local_root` 改为 `/Volumes/skillclaw-
 2. 经本地代理使用 agent。会话结束后（默认空闲超时为 180 秒），确认共享目录的 `team-lan/sessions/` 中出现会话文件；evolver 可能在下一轮处理时将其取走。
 3. 查看服务器日志和 `evolve_history.jsonl`。是否产生技能更新取决于证据质量，并非每轮必然更新。接受的变化写入 `team-lan/skills/`，客户端轮询后拉取到本地。
 
+如果只想同步技能、不导出会话和工具内容，运行 `skillclaw config sharing.upload_sessions false` 并重启代理。周期快照和会话结束上传都会关闭（包括本地共享目录），技能推送/拉取仍可用。该开关不删除已共享文件、不关闭本地记录，也不阻止向推理服务发送请求。首次 setup 会明确询问，默认不上传；已有配置保持原来的上传行为，需主动修改。
+
 同组用户共享会话证据和生成的技能。`group_id` 只是命名空间，不是权限隔离；不同信任组应使用独立共享根目录和文件系统/共享访问权限。本例没有自动隐私脱敏。默认发布模式为 `direct`；若需先让客户端验证候选技能，请使用上文的 `validated` 流程。
 
 Cloudflare R2 应使用**账号级 endpoint**，例如 `https://<ACCOUNT_ID>.r2.cloudflarestorage.com`，并配置 `sharing.backend s3`、`sharing.bucket <bucket-name>`、`sharing.region auto`。不要把 bucket 拼到 `sharing.endpoint` 后面；boto3 会单独发送 bucket。SkillClaw 会在访问存储前拒绝这种错误配置。参见 [R2 S3 API 文档](https://developers.cloudflare.com/r2/api/s3/api/)。

skillclaw/api_server.py

@@ -2117,7 +2117,7 @@ async def _close_session(self, session_id: str, reason: str = "explicit") -> Non
                 self.skill_manager._save_stats()
             turns = self._session_turns.pop(session_id, [])
             modified_skill_names = _extract_modified_skill_names(turns)
-            if turns and self.config.sharing_enabled:
+            if turns and self.config.sharing_enabled and self.config.sharing_upload_sessions:
                 self._safe_create_task(self._upload_session_data(session_id, turns))
             if self.config.sharing_enabled:
                 self._safe_create_task(self._pull_skills_from_cloud(skip_names=modified_skill_names))
@@ -3059,6 +3059,8 @@ async def _upload_session_data(
           - sessions: ``{group_id}/sessions/{session_id}.jsonl``
           - skills:   ``{group_id}/skills/{name}/SKILL.md``  (handled by SkillHub)
         """
+        if not self.config.sharing_enabled or not self.config.sharing_upload_sessions:
+            return False
         try:
             from .skill_hub import SkillHub
 
@@ -3103,7 +3105,7 @@ def _advance_user_turn_and_maybe_upload(self, session_id: str) -> int:
     def _maybe_upload_session_snapshot(self, session_id: str, user_turn_num: int) -> None:
         """Queue a session snapshot when the user-visible turn cadence is reached."""
         interval = max(0, int(getattr(self.config, "sharing_session_upload_interval", 0) or 0))
-        if not self.config.sharing_enabled or interval <= 0:
+        if not self.config.sharing_enabled or not self.config.sharing_upload_sessions or interval <= 0:
             return
         if user_turn_num <= 0 or user_turn_num % interval != 0:
             return

skillclaw/config.py

@@ -88,6 +88,7 @@ class SkillClawConfig:
     # Skill sharing (generic object storage)                              #
     # ------------------------------------------------------------------ #
     sharing_enabled: bool = False
+    sharing_upload_sessions: bool = True
     sharing_backend: str = ""
     sharing_endpoint: str = ""
     sharing_bucket: str = ""

skillclaw/config_store.py

@@ -77,6 +77,7 @@ def _resolve_config_file(config_file: Path | None = None) -> Path:
     },
     "sharing": {
         "enabled": False,
+        "upload_sessions": True,
         "backend": "",
         "endpoint": "",
         "bucket": "",
@@ -401,6 +402,7 @@ def to_skillclaw_config(self) -> SkillClawConfig:
             configure_openclaw=configure_openclaw,
             # Sharing
             sharing_enabled=bool(sharing.get("enabled", False)),
+            sharing_upload_sessions=bool(sharing.get("upload_sessions", True)),
             sharing_backend=sharing_backend,
             sharing_endpoint=sharing_endpoint,
             sharing_bucket=sharing_bucket,
@@ -536,6 +538,7 @@ def describe(self) -> str:
             lines += [
                 f"sharing.group:   {sharing.get('group_id', 'default')}",
                 f"sharing.alias:   {sharing.get('user_alias', '?')}",
+                f"sharing.upload_sessions: {sharing.get('upload_sessions', True)}",
                 f"sharing.auto_pull: {sharing.get('auto_pull_on_start', False)}",
                 "sharing.session_upload_interval: "
                 f"{_normalize_non_negative_int(sharing.get('session_upload_interval', 0), default=0)}",

skillclaw/setup_wizard.py

@@ -309,8 +309,13 @@ def run(self):
                 "Auto-pull shared skills on startup",
                 default=current_sharing.get("auto_pull_on_start", False),
             )
+            upload_sessions = _prompt_bool(
+                "Upload conversation and tool contents for shared skill evolution",
+                default=current_sharing.get("upload_sessions", current_sharing.get("enabled", False)),
+            )
             sharing_config = {
                 "enabled": True,
+                "upload_sessions": upload_sessions,
                 "backend": sharing_backend,
                 "group_id": group_id,
                 "user_alias": user_alias,

tests/test_session_upload_control.py

@@ -0,0 +1,51 @@
+from unittest.mock import AsyncMock, Mock
+
+import pytest
+
+from skillclaw.api_server import SkillClawAPIServer
+from skillclaw.config import SkillClawConfig
+from skillclaw.config_store import ConfigStore
+from skillclaw.skill_hub import SkillHub
+
+
+@pytest.mark.anyio
+@pytest.mark.parametrize("enabled,upload", [(True, False), (False, True)])
+async def test_disabled_upload_never_opens_storage_or_triggers_evolution(monkeypatch, enabled, upload):
+    server = object.__new__(SkillClawAPIServer)
+    server.config = SkillClawConfig(sharing_enabled=enabled, sharing_upload_sessions=upload)
+    storage = Mock(side_effect=AssertionError("must not open session storage"))
+    monkeypatch.setattr(SkillHub, "object_storage_from_config", storage)
+    server._trigger_evolve = AsyncMock()
+    assert await server._upload_session_data("private", [{"prompt_text": "private"}]) is False
+    await server._upload_session_snapshot_and_trigger("private", [{"prompt_text": "private"}])
+    storage.assert_not_called()
+    server._trigger_evolve.assert_not_awaited()
+
+
+def test_disabled_upload_never_schedules_snapshot():
+    server = object.__new__(SkillClawAPIServer)
+    server.config = SkillClawConfig(
+        sharing_enabled=True, sharing_upload_sessions=False, sharing_session_upload_interval=1
+    )
+    server._safe_create_task = Mock()
+    server._maybe_upload_session_snapshot("private", 1)
+    server._safe_create_task.assert_not_called()
+
+
+@pytest.mark.anyio
+async def test_skill_pull_still_works_with_session_upload_disabled(monkeypatch, tmp_path):
+    server = object.__new__(SkillClawAPIServer)
+    server.config = SkillClawConfig(sharing_enabled=True, sharing_upload_sessions=False, skills_dir=str(tmp_path))
+    server.skill_manager = None
+    hub = Mock()
+    hub.pull_skills.return_value = {"downloaded": 0, "skipped": 0, "failed": 0, "deleted": 0, "total_remote": 0}
+    monkeypatch.setattr(SkillHub, "from_config", lambda config: hub)
+    await server._pull_skills_from_cloud()
+    hub.pull_skills.assert_called_once()
+
+
+@pytest.mark.parametrize("setting,expected", [("", True), ("  upload_sessions: false\n", False)])
+def test_config_preserves_legacy_behavior_and_explicit_opt_out(tmp_path, setting, expected):
+    path = tmp_path / "config.yaml"
+    path.write_text("sharing:\n  enabled: true\n" + setting)
+    assert ConfigStore(path).to_skillclaw_config().sharing_upload_sessions is expected