fix(skills): keep load error text out of user-facing skill read 500s (#7155)

* fix(skills): keep load error text out of user-facing skill read 500s

GET /api/skills, /api/skills/custom and /api/skills/{name} are open to
every signed-in user, but their 500 detail echoed str(e). A malformed or
unreadable extensions_config.json (re-read on every load by the
user-scoped storage) made all three return the file's absolute host path,
plus the [Errno 13] text for a permission error.

Return a fixed detail like the diagnostics and reload routes already do,
chain the cause, and keep the full error in the Gateway log.

* docs(changelog): reference #7155 for the skill read 500 detail fix

* test(skills): pin capability skill discovery 500 through the real Gateway app

GET /api/capabilities/installations/skills reads the same user-scoped
storage as the skill routes but has no handler of its own. Its 500 body
is TraceMiddleware's fixed fallback, so drive create_app() with a
malformed extensions_config.json and assert no host path reaches the
body on any of the four reads.

CHANGELOG.md

@@ -6111,6 +6111,14 @@ This release closes that milestone with **621 merged pull requests**.
 
 ### Security
 
+- **skills:** `GET /api/skills`, `GET /api/skills/custom` and
+  `GET /api/skills/{name}`, which every signed-in user can call, no longer put
+  the raw exception text in their `500` detail. A malformed or unreadable
+  `extensions_config.json` made all three answer with the file's absolute host
+  path (and, for a permission error, the `[Errno 13]` message). They now return
+  a fixed `Failed to load skills.` / `Failed to list custom skills.` /
+  `Failed to get skill.`, and the full error stays in the Gateway log. ([#7155])
+
 - **skillscan:** The pre-LLM static gate now matches fork bombs by shape
   instead of only the canonical `:(){ :|:& };:` spelling. `resource-fork-bomb`
   and the fork-bomb branch of `shell-destructive-command` matched just that
@@ -11911,3 +11919,4 @@ with **180 merged pull requests** since the first 2.0 milestone tag.
 [#7099]: https://github.com/bytedance/deer-flow/pull/7099
 [#7104]: https://github.com/bytedance/deer-flow/pull/7104
 [#7123]: https://github.com/bytedance/deer-flow/pull/7123
+[#7155]: https://github.com/bytedance/deer-flow/pull/7155

CHANGELOG_zh.md

@@ -4606,6 +4606,13 @@
 
 ### 安全
 
+- **技能：** 所有已登录用户都能调用的 `GET /api/skills`、`GET /api/skills/custom`
+  和 `GET /api/skills/{name}`，其 `500` 响应的 detail 不再包含原始异常文本。此前
+  `extensions_config.json` 格式错误或不可读时，这三个接口都会在响应中返回该文件的
+  宿主机绝对路径（权限错误时还会附带 `[Errno 13]` 信息）。现在它们分别返回固定的
+  `Failed to load skills.` / `Failed to list custom skills.` /
+  `Failed to get skill.`，完整错误只记录在 Gateway 日志中。 ([#7155])
+
 - **skillscan：** LLM 之前的静态门禁现在按形状匹配 fork 炸弹，而不再只匹配
   `:(){ :|:& };:` 这一规范写法。`resource-fork-bomb` 和 `shell-destructive-command`
   的 fork 炸弹分支此前只匹配那一种写法，因此 #6651 报告的改名或改空格的载荷——
@@ -9463,3 +9470,4 @@ DeerFlow 2.0 是围绕"超级智能体"框架的彻底重写，核心包含子
 [#7099]: https://github.com/bytedance/deer-flow/pull/7099
 [#7104]: https://github.com/bytedance/deer-flow/pull/7104
 [#7123]: https://github.com/bytedance/deer-flow/pull/7123
+[#7155]: https://github.com/bytedance/deer-flow/pull/7155

backend/app/gateway/routers/AGENTS.md

@@ -16,6 +16,11 @@ create-only: collisions return 409, never overwrite or upsert. Ordinary creation
 and import share `_persist_new_agent`, which uses `run_drained_write` so a
 cancelled request still drains the owned persistence worker.
 
+`GET /api/skills`, `/api/skills/custom` and `/api/skills/{name}` serve every
+signed-in user, so their unexpected-failure 500s return a fixed detail. Load
+errors (e.g. a malformed `extensions_config.json`) carry host paths and stay in
+the server log.
+
 Skill mutation operation lists return `next_cursor: null` when `has_more` is false, including empty pages after a supplied `after_id`. The next cursor is the final returned operation only when a further page exists.
 
 `GET /api/agent-tool-groups` projects only configured group names for the

backend/app/gateway/routers/skills.py

@@ -508,7 +508,7 @@ async def list_skills(request: Request, config: AppConfig = Depends(get_config))
         skills = await asyncio.to_thread(_load_user_skills, config)
     except Exception as e:
         logger.error(f"Failed to load skills: {e}", exc_info=True)
-        raise HTTPException(status_code=500, detail=f"Failed to load skills: {str(e)}")
+        raise HTTPException(status_code=500, detail="Failed to load skills.") from e
 
     visible_skills = await _filter_visible_skills(request, config, skills)
 
@@ -655,7 +655,7 @@ async def list_custom_skills(request: Request, config: AppConfig = Depends(get_c
         return SkillsListResponse(skills=[_skill_to_response(skill) for skill in visible_skills])
     except Exception as e:
         logger.error("Failed to list custom skills: %s", e, exc_info=True)
-        raise HTTPException(status_code=500, detail=f"Failed to list custom skills: {str(e)}")
+        raise HTTPException(status_code=500, detail="Failed to list custom skills.") from e
 
 
 @router.get("/skills/diagnostics/custom", response_model=SkillDiagnosticsResponse, summary="List Owned Custom Skill YAML Errors")
@@ -1027,7 +1027,7 @@ async def get_skill(skill_name: str, request: Request, config: AppConfig = Depen
         raise
     except Exception as e:
         logger.error(f"Failed to get skill {skill_name}: {e}", exc_info=True)
-        raise HTTPException(status_code=500, detail=f"Failed to get skill: {str(e)}")
+        raise HTTPException(status_code=500, detail="Failed to get skill.") from e
 
 
 def _write_extensions_skill_state(

backend/tests/test_skills_custom_router.py

@@ -31,6 +31,12 @@ def _make_admin_user() -> User:
     return User(email="admin-test@example.com", password_hash="x", system_role="admin", id=uuid4())
 
 
+def _make_member_user() -> User:
+    from uuid import uuid4
+
+    return User(email="member-test@example.com", password_hash="x", system_role="user", id=uuid4())
+
+
 def _skill_content(name: str, description: str = "Demo skill") -> str:
     return f"---\nname: {name}\ndescription: {description}\n---\n\n# {name}\n"
 
@@ -778,7 +784,7 @@ def test_custom_skill_delete_preserves_history_and_allows_restore(monkeypatch, t
         assert refresh_calls == [("refresh", "default"), ("refresh", "default")]
 
 
-def _make_demo_custom_skill_app(monkeypatch, tmp_path: Path) -> FastAPI:
+def _patch_demo_custom_skill_config(monkeypatch, tmp_path: Path) -> SimpleNamespace:
     from deerflow.config.paths import Paths
 
     custom_dir = _user_custom_dir(tmp_path, "default") / "demo-skill"
@@ -793,7 +799,11 @@ def _make_demo_custom_skill_app(monkeypatch, tmp_path: Path) -> FastAPI:
     monkeypatch.setattr("deerflow.config.paths.get_paths", lambda: Paths(base_dir=tmp_path))
     monkeypatch.setattr("deerflow.config.paths._paths", None)
     monkeypatch.setattr("app.gateway.routers.skills.get_effective_user_id", lambda: "default")
-    return _make_test_app(config)
+    return config
+
+
+def _make_demo_custom_skill_app(monkeypatch, tmp_path: Path, *, user_factory=None) -> FastAPI:
+    return _make_test_app(_patch_demo_custom_skill_config(monkeypatch, tmp_path), user_factory=user_factory)
 
 
 @pytest.mark.parametrize(
@@ -859,6 +869,69 @@ def test_custom_skill_corrupt_history_record_stays_server_error(monkeypatch, tmp
     assert response.json()["detail"].startswith(detail)
 
 
+@pytest.mark.parametrize(
+    ("path", "detail"),
+    [
+        ("/api/skills", "Failed to load skills."),
+        ("/api/skills/custom", "Failed to list custom skills."),
+        ("/api/skills/demo-skill", "Failed to get skill."),
+    ],
+)
+def test_member_skill_reads_keep_load_error_text_out_of_500(monkeypatch, tmp_path, caplog, path, detail):
+    """Any signed-in user reaches these reads, so a load failure's host path stays in the server log."""
+    import logging
+
+    extensions_config = tmp_path / "extensions_config.json"
+    extensions_config.write_text("{not json", encoding="utf-8")
+    monkeypatch.setenv("DEER_FLOW_EXTENSIONS_CONFIG_PATH", str(extensions_config))
+    app = _make_demo_custom_skill_app(monkeypatch, tmp_path, user_factory=_make_member_user)
+
+    with TestClient(app) as client, caplog.at_level(logging.ERROR, logger=skills_router.__name__):
+        response = client.get(path)
+
+    assert response.status_code == 500
+    assert response.json() == {"detail": detail}
+    assert str(extensions_config) in caplog.text
+
+
+@pytest.mark.parametrize(
+    "path",
+    [
+        "/api/skills",
+        "/api/skills/custom",
+        "/api/skills/demo-skill",
+        "/api/capabilities/installations/skills",
+    ],
+)
+def test_gateway_skill_reads_keep_load_error_text_out_of_500(monkeypatch, tmp_path, path):
+    """Drive the real Gateway app: capability discovery has no handler of its own, so its 500 body is TraceMiddleware's fixed fallback until a handler for the raised class or a changed fallback echoes the exception."""
+    import app.gateway.app as app_module
+    from deerflow.extensions import reset_loaded_extensions, reset_runtime_diagnostics
+
+    extensions_config = tmp_path / "extensions_config.json"
+    extensions_config.write_text("{not json", encoding="utf-8")
+    monkeypatch.setenv("DEER_FLOW_EXTENSIONS_CONFIG_PATH", str(extensions_config))
+    monkeypatch.setenv("DEER_FLOW_AUTH_DISABLED", "1")
+    config = _patch_demo_custom_skill_config(monkeypatch, tmp_path)
+    config.authorization = AuthorizationConfig(enabled=False)
+
+    def _no_config_yaml():
+        raise FileNotFoundError("config.yaml")
+
+    # create_app() reads only the plugin list from config.yaml; keep a local one out of the test.
+    monkeypatch.setattr(app_module, "get_app_config", _no_config_yaml)
+    try:
+        gateway = app_module.create_app()
+        gateway.dependency_overrides[get_config] = lambda: config
+        response = TestClient(gateway, raise_server_exceptions=False).get(path)
+    finally:
+        reset_loaded_extensions()
+        reset_runtime_diagnostics()
+
+    assert response.status_code == 500
+    assert str(tmp_path) not in response.text
+
+
 def test_custom_skill_delete_continues_when_history_write_is_readonly(monkeypatch, tmp_path):
     skills_root = tmp_path / "skills"
     from deerflow.config.paths import Paths