feat: 扩展后端 MCP 应用运行能力 + 加封装服务 + 外部agent测试结果 - #2460
Merged
Merged
Conversation
Contributor
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
📝 WalkthroughWalkthrough本次改动扩展应用与自定义 operation 运行能力，新增 HTTP/MCP 服务接口、MCP 使用指南和 GUI 本机服务管理页，并同步更新后端文档及测试仓库检出回退流程。 Changes后端运行与服务能力扩展
Estimated code review effort: 4 (Complex) | ~60 minutes Sequence Diagram(s)sequenceDiagram
participant MCPClient
participant MCPService
participant OperationRegistry
participant ZzzBackendContext
participant RunSlot
MCPClient->>MCPService: run_operation(op_id, args, block)
MCPService->>OperationRegistry: resolve and validate operation
MCPService->>ZzzBackendContext: submit operation
ZzzBackendContext->>RunSlot: _start(op_factory)
RunSlot-->>MCPService: started status or final result
MCPService-->>MCPClient: structured response
sequenceDiagram
participant GUI
participant McpServiceInterface
participant McpServiceRunner
participant ServerProcess
GUI->>McpServiceInterface: start/stop/restart
McpServiceInterface->>McpServiceRunner: execute action
McpServiceRunner->>ServerProcess: launch or terminate process
McpServiceRunner->>ServerProcess: probe /health
ServerProcess-->>McpServiceRunner: health response
McpServiceRunner-->>McpServiceInterface: result and status
Possibly related PRs
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 2
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (2)
src/one_dragon/base/operation/application/application_run_context.py (1)
433-449: 🎯 Functional Correctness | 🟠 Major | ⚡ Quick win
app.execute()抛异常时未设置last_application_result，会被下游误判为成功。
except Exception:分支只记录日志，self.last_application_result保持在函数开头被清空的None。而backend_context.py的ApplicationRunSlot._run_application读取到result is None时会走elif result is None:分支，将其固化为OperationResult(success=True, status='应用运行结束')——把真实的执行异常报告为运行成功。对比
RunSlot._run()（同文件所在改动集）在except Exception as e:分支中会显式构造OperationResult(success=False, status='执行异常')，这里应保持一致语义。🐛 建议修复：异常路径显式落地失败结果
- except Exception: - log.error("运行应用 {} 失败", app_id, exc_info=True) + except Exception as e: + log.error("运行应用 {} 失败", app_id, exc_info=True) + self.last_application_result = OperationResult(success=False, status=f'执行异常: {e}')🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/one_dragon/base/operation/application/application_run_context.py` around lines 433 - 449, `application_run_context.py` 中 `app.execute()` 的异常分支没有为 `last_application_result` 赋失败结果，导致下游把异常误判为成功；请在 `run_application` 的 `except Exception` 分支里为 `self.last_application_result` 显式设置一个失败的 `OperationResult`，并保持与同文件中的 `RunSlot._run()` 异常处理语义一致，确保 `backend_context.py` 读取时不会把 `None` 当作成功结束。src/zzz_od/backend/backend_context.py (1)
760-805: 🩺 Stability & Availability | 🔴 Critical | 🏗️ Heavy lift
basic_run_slot与app_run_slot之间的启动互斥检查存在 TOCTOU 竞态，可能同时运行。
start_run检查self.app_run_slot.is_running()，run_one_dragon/run_standalone_app检查self.basic_run_slot.is_running()，但“检查”与“启动”分别持有两个槽各自独立的锁，不是一次原子操作。并发场景下（例如 MCP/HTTP 几乎同时发起start_run与run_one_dragon），两次检查都可能读到“未运行”，随后两个槽各自成功启动，导致一条龙应用运行与普通 operation 同时抢占游戏控制器，破坏代码里反复强调的“单跑道”设计（如RunSlot/ApplicationRunSlot文档所述）。🔒 建议修复：引入跨槽的启动锁
class ZzzBackendContext: def __init__(self, ctx: ZContext) -> None: self._ctx: ZContext = ctx + self._start_lock: threading.Lock = threading.Lock() self.basic_run_slot: RunSlot = RunSlot(ctx) self.run_slot: RunSlot = self.basic_run_slot self.app_run_slot: ApplicationRunSlot = ApplicationRunSlot(ctx) def start_run(self, source, op_factory): - if self.app_run_slot.is_running(): - return False, None - return self.basic_run_slot._start_run(source, op_factory) + with self._start_lock: + if self.app_run_slot.is_running(): + return False, None + return self.basic_run_slot._start_run(source, op_factory) def run_one_dragon(self, source: str): self._ensure_ready() - if self.basic_run_slot.is_running(): - return False, None - self._refresh_runtime_config() - return self.app_run_slot._start_application(...) + with self._start_lock: + if self.basic_run_slot.is_running(): + return False, None + self._refresh_runtime_config() + return self.app_run_slot._start_application(...)
run_standalone_app同理需要包在同一把_start_lock里。🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/zzz_od/backend/backend_context.py` around lines 760 - 805, `start_run`, `run_one_dragon`, and `run_standalone_app` have a TOCTOU race because they check `basic_run_slot` and `app_run_slot` separately before starting. Fix this by adding a shared cross-slot start lock or a single atomic arbitration point so the “is running” check and `_start_run`/`_start_application` launch happen under the same critical section. Use the existing `RunSlot`/`ApplicationRunSlot` start path and ensure all three methods coordinate through the same lock to preserve single-run exclusivity.
🧹 Nitpick comments (3)
src/zzz_od/gui/view/devtools/mcp_service_interface.py (3)
218-222: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win
__init__不应标注-> None。
McpServiceRunner.__init__和McpServiceInterface.__init__都添加了-> None返回类型标注，这与项目既有约定不一致。Based on learnings, "follow the established style convention for
__init__methods: do not add a-> Nonereturn type annotation to__init__signatures" was previously flagged in this codebase.♻️ 建议修改
- def __init__(self, action: str, port: int, parent: QWidget | None = None) -> None: + def __init__(self, action: str, port: int, parent: QWidget | None = None):- def __init__(self, ctx: ZContext, parent: QWidget | None = None) -> None: + def __init__(self, ctx: ZContext, parent: QWidget | None = None):Also applies to: 245-261
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/zzz_od/gui/view/devtools/mcp_service_interface.py` around lines 218 - 222, Remove the explicit return type annotation from the affected __init__ methods to match the project’s established style. Update both McpServiceRunner.__init__ and McpServiceInterface.__init__ so their signatures do not include -> None, while keeping the existing parameters and initialization logic unchanged.Source: Learnings
40-40: 📐 Maintainability & Code Quality | 🔵 Trivial | 🏗️ Heavy lift端口配置硬编码，未落到 YAML/YamlConfig。
DEFAULT_MCP_PORT和端口输入框只是页面内的临时状态，每次进入页面都会重置为23001（self.port_card.setValue(str(DEFAULT_MCP_PORT), emit_signal=False)），用户修改后的端口不会被持久化。As per coding guidelines, "配置改动优先落到 YAML 与对应的
YamlConfig子类，不要随意散落硬编码配置。" 建议把端口做成一个YamlConfig字段，页面读取/写入该配置而不是每次都回退到硬编码默认值。Also applies to: 277-279
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/zzz_od/gui/view/devtools/mcp_service_interface.py` at line 40, The MCP port is hardcoded as DEFAULT_MCP_PORT and only stored in the view state, so user changes are lost when reopening the page. Move the port setting into the relevant YamlConfig subclass and have the devtools MCP UI read and write that config instead of resetting via self.port_card.setValue(str(DEFAULT_MCP_PORT), emit_signal=False); update the persistence flow around the port handling logic in the MCP service interface so the chosen value survives page refreshes.Source: Coding guidelines
315-323: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win
PlainTextEdit直接通过addSettingCard加入SettingCardGroup，可能不符合 Fluent 卡片布局预期。
SettingCardGroup.addSettingCard设计目标是容纳标准SettingCard（固定高度、带标题/描述），这里塞入一个setMinimumHeight(180)的日志文本框，虽然类型上兼容（该方法接受任意QWidget），但视觉呈现上可能与其余卡片风格不一致。As per coding guidelines, "GUI 代码优先复用
pyside6-fluent-widgets与现有项目组件，保持 Fluent Design。" 建议确认实际渲染效果，或考虑用项目里已有的日志/文本展示容器组件包一层，而不是把PlainTextEdit直接当作 SettingCard 加入分组。🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/zzz_od/gui/view/devtools/mcp_service_interface.py` around lines 315 - 323, The MCP service view currently adds a PlainTextEdit directly into SettingCardGroup via addSettingCard, which may clash with the expected Fluent card layout. Update the devtools UI in mcp_service_interface by either wrapping the log view in an existing project-style container or using a proper SettingCard-style component, and keep the message_box setup aligned with the rest of the Fluent design. Verify the rendered appearance still fits the surrounding SettingCardGroup and content.add_widget structure.Source: Coding guidelines
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@src/zzz_od/backend/mcp/service_app.py`:
- Around line 92-97: make_list_applications 直接返回
backend.list_applications()，缺少与同层工具一致的异常兜底；请在 list_applications 里补上 try/except，像
make_run_one_dragon、make_run_standalone_app 一样把 BackendNotReadyError
和其它异常转换为结构化错误结果，而不是让异常直接冒泡到 MCP 框架。优先参考
make_list_applications、make_run_one_dragon 和 make_run_standalone_app
这几个符号来统一处理方式，并保持“无副作用”的正常返回路径不变。
In `@src/zzz_od/gui/view/devtools/mcp_service_interface.py`:
- Around line 333-337: The on_interface_hidden cleanup in
mcp_service_interface.py stops the timers but does not wait for self._runner to
finish, so the worker thread can still be running during page hide or
destruction. Update on_interface_hidden, or the shared teardown path it uses, to
stop the timers and then wait for self._runner to exit, using a timeout if
needed, so the runner is not destroyed while start/restart work is still in
progress.
---
Outside diff comments:
In `@src/one_dragon/base/operation/application/application_run_context.py`:
- Around line 433-449: `application_run_context.py` 中 `app.execute()` 的异常分支没有为
`last_application_result` 赋失败结果，导致下游把异常误判为成功；请在 `run_application` 的 `except
Exception` 分支里为 `self.last_application_result` 显式设置一个失败的
`OperationResult`，并保持与同文件中的 `RunSlot._run()` 异常处理语义一致，确保 `backend_context.py`
读取时不会把 `None` 当作成功结束。
In `@src/zzz_od/backend/backend_context.py`:
- Around line 760-805: `start_run`, `run_one_dragon`, and `run_standalone_app`
have a TOCTOU race because they check `basic_run_slot` and `app_run_slot`
separately before starting. Fix this by adding a shared cross-slot start lock or
a single atomic arbitration point so the “is running” check and
`_start_run`/`_start_application` launch happen under the same critical section.
Use the existing `RunSlot`/`ApplicationRunSlot` start path and ensure all three
methods coordinate through the same lock to preserve single-run exclusivity.
---
Nitpick comments:
In `@src/zzz_od/gui/view/devtools/mcp_service_interface.py`:
- Around line 218-222: Remove the explicit return type annotation from the
affected __init__ methods to match the project’s established style. Update both
McpServiceRunner.__init__ and McpServiceInterface.__init__ so their signatures
do not include -> None, while keeping the existing parameters and initialization
logic unchanged.
- Line 40: The MCP port is hardcoded as DEFAULT_MCP_PORT and only stored in the
view state, so user changes are lost when reopening the page. Move the port
setting into the relevant YamlConfig subclass and have the devtools MCP UI read
and write that config instead of resetting via
self.port_card.setValue(str(DEFAULT_MCP_PORT), emit_signal=False); update the
persistence flow around the port handling logic in the MCP service interface so
the chosen value survives page refreshes.
- Around line 315-323: The MCP service view currently adds a PlainTextEdit
directly into SettingCardGroup via addSettingCard, which may clash with the
expected Fluent card layout. Update the devtools UI in mcp_service_interface by
either wrapping the log view in an existing project-style container or using a
proper SettingCard-style component, and keep the message_box setup aligned with
the rest of the Fluent design. Verify the rendered appearance still fits the
surrounding SettingCardGroup and content.add_widget structure.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: 568a7e67-cf0f-4fa5-985c-c13761cbf999
📒 Files selected for processing (18)
docs/develop/zzz/backend/README.mddocs/develop/zzz/backend/architecture.mddocs/develop/zzz/backend/entry.mddocs/develop/zzz/backend/http.mddocs/develop/zzz/backend/mcp.mdsrc/one_dragon/base/operation/application/application_factory.pysrc/one_dragon/base/operation/application/application_run_context.pysrc/one_dragon/base/push/push_service.pysrc/zzz_od/backend/backend_context.pysrc/zzz_od/backend/entry/server.pysrc/zzz_od/backend/http/routes.pysrc/zzz_od/backend/http/service_routes.pysrc/zzz_od/backend/mcp/app.pysrc/zzz_od/backend/mcp/prompts.pysrc/zzz_od/backend/mcp/service_app.pysrc/zzz_od/backend/schemas.pysrc/zzz_od/gui/view/devtools/app_devtools_interface.pysrc/zzz_od/gui/view/devtools/mcp_service_interface.py
Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: glm-5.2 <noreply@bigmodel.cn>
run_application 的 except 分支此前只 log 不写 last_application_result， 导致 app.execute() 抛异常时该字段保持初始 None，backend 运行槽可能将 异常误判为成功。现将 OperationResult 提为运行时导入，并在 except 分支 固化 success=False 的失败终态。 Task 1 of 后端运行槽合并 + 自定义 operation 运行入口 Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: glm-5.2 <noreply@bigmodel.cn>
RunSlot 新增 RunType(APPLICATION/OPERATION)枚举与 op_id/run_type 字段; _start 替代 _start_run(op_factory 与 app_id 互斥、check+submit 同锁原子), _run 顶层 try/except/finally 固化终态并按 app/op 分派(app 委托 run_application、 op 槽自管 start_running/execute/stop_running),_node_name 统一读 current_op or current_application。ZzzBackendContext.start_run 改调 run_slot._start 并透传 display_name,保证 open_game 路径中间态可跑。 暂不删 ApplicationRunSlot / 不动 query_status/stop(Task 3 收敛)。 Task 2 of 后端运行槽合并 + 自定义 operation 运行入口 Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: glm-5.2 <noreply@bigmodel.cn>
- 删 ApplicationRunSlot 整个类(约 -130 行)及 app_run_slot/basic_run_slot 双字段 - __init__ 只留 self.run_slot: RunSlot = RunSlot(ctx)(无 host) - 新增 _start_app(source, app_id, group_id):app 路径统一入口, refresh_config 作钩子注入槽线程(拒绝路径不刷新,修刷新竞态) - run_one_dragon/run_standalone_app 改走 _start_app - start_run 删 app_run_slot.is_running() 守卫(互斥由 _start 锁内原子保证) - query_status/stop 塌缩为单次 run_slot._query_status()/_stop()(删 6 分支跨槽仲裁) - list_applications 删 _refresh_runtime_config()(只读路径无副作用) - 删不再使用的 Application TYPE_CHECKING 导入 Task 3 of backend run-slot 统一设计(spec §4.3/§4.2) Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: glm-5.2 <noreply@bigmodel.cn>
…erationInfo/OperationListResult) Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: glm-5.2 <noreply@bigmodel.cn>
- service_app: list_operations / describe_operation / run_operation 三 tool 工厂
run_operation 经 resolve_op_class + validate_args 校验后,op_factory 闭包 bake args
提交 run_slot._start;并发拒绝/校验失败/异常一律返 {started: False, error}
- service_routes: GET /game/operations、GET /game/operations/describe?op_id=、
POST /game/run/operation?op_id=&block=(args 走 JSON body);业务失败一律 200+body
- app.py: 注册三个新 MCP tool
Co-Authored-By: Claude Code <noreply@anthropic.com>
Co-Authored-By: glm-5.2 <noreply@bigmodel.cn>
Task 7 文档更新（单槽分派模型落地后的主 spec 同步）: - architecture.md: 删除 ApplicationRunSlot/双槽描述，改为单 RunSlot + app/op 分派 (app 委托 run_application、op 自管 start_running/execute/stop_running;互斥收进 _start 锁内;字段 op_id/run_type/app/current_op);新增 operation_registry 小节 - mcp.md: 工具表加 list_operations/describe_operation/run_operation(op_id 格式、 args 闭包 bake、block),说明 run_operation 是通用 op 运行入口;计数订正为 19 - http.md: 加 GET /game/operations、/game/operations/describe、 POST /game/run/operation(op_id 走 query、args 走 body、业务失败 200+body) - entry.md: 新增路由总览表(原有 + 自定义 op 3 端点) - README.md: mermaid 单槽化 + 已实现列表补自定义 op 运行 Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: glm-5.2 <noreply@bigmodel.cn>
Contributor
There was a problem hiding this comment.
🧹 Nitpick comments (1)
src/one_dragon/base/operation/application/application_run_context.py (1)
440-447: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value建议添加
# noqa: BLE001以与backend_context.py保持一致。此处的
except Exception是有意义的异常处理（固化失败终态），Ruff 的 BLE001 警告属于误报。backend_context.py中相同模式已使用# noqa: BLE001抑制，建议此处保持一致。注：RUF003 对第 444 行中文逗号的警告不予处理，符合路径指令中"不要建议把中文标点符号改为英文"的要求。
🔧 建议的修改
op_result = app.execute() # run_application 的布尔返回值只表示是否成功启动；详细执行结果由这里保留。 self.last_application_result = op_result - except Exception as e: + except Exception as e: # noqa: BLE001 log.error("运行应用 {} 失败", app_id, exc_info=True)🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/one_dragon/base/operation/application/application_run_context.py` around lines 440 - 447, 在应用执行异常处理的 except Exception 语句上添加 # noqa: BLE001，参照 backend_context.py 的相同模式，保留现有的失败终态固化与日志处理逻辑不变。Source: Linters/SAST tools
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In `@src/one_dragon/base/operation/application/application_run_context.py`:
- Around line 440-447: 在应用执行异常处理的 except Exception 语句上添加 # noqa: BLE001，参照
backend_context.py 的相同模式，保留现有的失败终态固化与日志处理逻辑不变。
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: cb9a328b-389c-4f9d-8064-52d28c7f99aa
📒 Files selected for processing (12)
docs/develop/zzz/backend/README.mddocs/develop/zzz/backend/architecture.mddocs/develop/zzz/backend/entry.mddocs/develop/zzz/backend/http.mddocs/develop/zzz/backend/mcp.mdsrc/one_dragon/base/operation/application/application_run_context.pysrc/zzz_od/backend/backend_context.pysrc/zzz_od/backend/http/service_routes.pysrc/zzz_od/backend/mcp/app.pysrc/zzz_od/backend/mcp/service_app.pysrc/zzz_od/backend/operation_registry.pysrc/zzz_od/backend/schemas.py
✅ Files skipped from review due to trivial changes (2)
- docs/develop/zzz/backend/README.md
- docs/develop/zzz/backend/http.md
🚧 Files skipped from review as they are similar to previous changes (2)
- src/zzz_od/backend/mcp/app.py
- docs/develop/zzz/backend/entry.md
…main) cross-repo PR 贡献者没 fork 测试仓时,原逻辑直接掉到一条龙 main(可能与 PR 代码不匹配)。 新增中间级:一条龙测试仓同名分支(mcp-service),让 maintainer 在主测试仓 push 同名测试即可被 CI 取到;都没有再 main 兜底。 Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: glm-5.2 <noreply@bigmodel.cn>
Contributor
There was a problem hiding this comment.
🧹 Nitpick comments (1)
.github/workflows/test-check.yml (1)
34-42: 🔒 Security & Privacy | 🔵 Trivial | ⚡ Quick win三个 checkout 步骤均未设置
persist-credentials: false，存在 GITHUB_TOKEN 泄露风险。
actions/checkout@v5默认将 GITHUB_TOKEN 持久化到.git/config。对于 fork 检出步骤（Line 38），贡献者 fork 的测试代码会在 workflow 中执行，恶意测试代码可从zzz-od-test/.git/config读取并外泄 token。这是已知的 artipacked 漏洞模式。建议在所有三个 checkout 步骤的
with块中添加persist-credentials: false：🔒️ 建议的修复
- name: Checkout test repo (fork, same branch) id: checkout-test-repo uses: actions/checkout@v5 with: repository: ${{ github.event_name == 'pull_request' && (github.event.pull_request.head.repo.full_name && format('{0}/zzz-od-test', github.event.pull_request.head.repo.owner.login) || 'OneDragon-Anything/zzz-od-test') || 'OneDragon-Anything/zzz-od-test' }} path: zzz-od-test ref: ${{ github.event_name == 'pull_request' && github.head_ref || 'main' }} fetch-depth: 1 + persist-credentials: false continue-on-error: true - name: Checkout test repo (upstream, same branch) id: checkout-test-repo-upstream if: steps.checkout-test-repo.outcome == 'failure' uses: actions/checkout@v5 with: repository: OneDragon-Anything/zzz-od-test path: zzz-od-test ref: ${{ github.head_ref || 'main' }} fetch-depth: 1 + persist-credentials: false continue-on-error: true - name: Checkout test repo (upstream, main) id: checkout-test-repo-default if: steps.checkout-test-repo-upstream.outcome == 'failure' uses: actions/checkout@v5 with: repository: OneDragon-Anything/zzz-od-test path: zzz-od-test ref: main fetch-depth: 1 + persist-credentials: falseAlso applies to: 44-53, 55-63
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.github/workflows/test-check.yml around lines 34 - 42, Update all three actions/checkout@v5 steps, including “Checkout test repo (fork, same branch)” and the two additional checkout steps, to set persist-credentials to false within each with block; preserve their existing repository, path, ref, and fetch-depth settings.Source: Linters/SAST tools
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In @.github/workflows/test-check.yml:
- Around line 34-42: Update all three actions/checkout@v5 steps, including
“Checkout test repo (fork, same branch)” and the two additional checkout steps,
to set persist-credentials to false within each with block; preserve their
existing repository, path, ref, and fetch-depth settings.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: c4bc1278-4972-4b81-99f3-96798dcb6fdb
📒 Files selected for processing (1)
.github/workflows/test-check.yml
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
新增和修改内容
MCP / HTTP 后端能力
run_one_dragon(block=False)run_standalone_app(app_id=None, block=False)list_applicationsGET /healthGET /game/applicationsPOST /game/run/one-dragonPOST /game/run/standaloneopen_game(enter=True, block=True)analyze_screen(screenshot=None, save_image=False)click_gameinput_text运行状态与运行槽
ApplicationRunSlot(继承RunSlot)，用于「一条龙」和 「独立应用 」运行。RunSlot作为基础 operation 运行槽。ZzzBackendContext.query_status()/stop()统一覆盖基础 operation 与应用运行。run_context.run_application，避免 MCP/HTTP 与 GUI 应用运行逻辑分叉。MCP prompts 与帮助工具
zzz_check_statuszzz_run_one_dragonzzz_run_standalone_applist_mcp_usage_guidesget_mcp_usage_guideGUI
文档
/health探测和测试仓配合方式。测试
本地已执行：
新界面示意
Codex调用结果
Summary by CodeRabbit
/health；扩展 HTTP/game/*：应用列表、应用运行（阻塞/非阻塞）、以及自定义 operation 列表/描述/运行；/game/close未就绪返回 503。