Add a /debug route to NAT MCP frontend to expose MCP tools #813
Conversation
Signed-off-by: Daniel Wang <daniewang@nvidia.com>
zhongxuanwang-nv
added
feature request
WalkthroughAdds a new public debug HTTP endpoint (/debug/tools/list) to the MCP front-end for listing registered tools and optionally their input schemas; implements the handler in MCPFrontEndPluginWorkerBase, wires it into routing, adds tests covering many query combinations, and documents the endpoint. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor Client
participant HTTP as HTTP Server
participant Worker as MCPFrontEndPluginWorkerBase
participant MCP as FastMCP
note over Worker,MCP: Endpoint: GET /debug/tools/list\nQuery: name (repeat/comma), detail=true|false
Client->>HTTP: GET /debug/tools/list?name=toolA&detail=true
HTTP->>Worker: invoke route handler
Worker->>MCP: retrieve registered functions/tools
alt name provided
Worker->>Worker: parse `name` values (split, dedupe)
alt any requested missing
Worker-->>HTTP: 404 Not Found (missing names)
else
opt include schemas
Worker->>Worker: build schema per tool (Pydantic model_json_schema or ChatRequest summary)
end
Worker-->>HTTP: 200 OK (server_name, count, tools[] or single tool)
end
else no name provided
alt detail=true
Worker->>Worker: include schemas for all tools
else
Worker->>Worker: return simplified list (no schemas)
end
Worker-->>HTTP: 200 OK (server_name, count, tools[])
end
HTTP-->>Client: JSON response
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Pre-merge checks and finishing touches❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✨ Finishing touches
🧪 Generate unit tests
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 review |
|
@zhongxuanwang-nv I'll review the changes in this PR that adds the /debug route to the NAT MCP frontend for exposing MCP tools and their schema. 🧠 Learnings used✅ Actions performedReview triggered.
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py (1)
270-272: Gate /debug endpoints behind an explicit opt‑in flag.Avoid exposing internals by default when bound to 0.0.0.0. Make the route opt‑in via an env flag.
Apply this diff:
- # After registration, expose debug endpoints for tool/schema inspection - self._setup_debug_endpoints(mcp, functions) + # After registration, optionally expose debug endpoints for tool/schema inspection + if os.getenv("NAT_MCP_ENABLE_DEBUG_ROUTES", "").strip().lower() in {"1", "true", "yes", "on"}: + logger.info("Debug endpoints enabled via NAT_MCP_ENABLE_DEBUG_ROUTES.") + self._setup_debug_endpoints(mcp, functions) + else: + logger.info("Debug endpoints disabled. Set NAT_MCP_ENABLE_DEBUG_ROUTES=1 to enable.")Also add the import:
import logging +import os
🧹 Nitpick comments (7)
docs/source/workflows/mcp/mcp-server.md (1)
101-116: Show error/404 behavior and detail toggles for completeness.Add a small example of the 404 response and one “detail=false” example with names to match implementation.
Apply this doc tweak:
# ...existing examples... You can control the amount of detail using the `detail` query parameter: - When requesting specific tool(s) with `name`, detailed schema is returned by default. Pass `detail=false` to suppress schemas: ```bash curl -s "http://localhost:9901/debug/tools/list?name=calculator_multiply&detail=false" | jq
+If any requested tool is missing, a 404 is returned:
+
+bash +curl -s "http://localhost:9901/debug/tools/list?name=not_a_tool" | jq +# { +# "error": "Tool(s) not found: not_a_tool", +# "available_tools": ["calculator_multiply", "..."] +# } +</blockquote></details> <details> <summary>src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py (6)</summary><blockquote> `106-112`: **Fix docstring and add return type.** Docstring implies single tool; code supports multiple. Add return type for compliance. Apply this diff: ```diff - def _setup_debug_endpoints(self, mcp: FastMCP, functions: dict[str, Function]): + def _setup_debug_endpoints(self, mcp: FastMCP, functions: dict[str, Function]) -> None: """Set up HTTP debug endpoints for introspecting tools and schemas. - Exposes: - - GET /debug/tools/list: List tools. Optional query param `name` returns details for a single tool. + Exposes: + - GET /debug/tools/list: List tools. Optional query param `name` (one or more) returns details for the selected tool(s). """
121-131: Deduplicate requested tool names while preserving order.Avoid repeated entries when
nameis provided multiple times.Apply this diff:
for raw in names_param_list: # if p.strip() is empty, it won't be included in the list! parts = [p.strip() for p in raw.split(",") if p.strip()] names.extend(parts) + # Deduplicate while preserving the first occurrence order + names = list(dict.fromkeys(names)) detail_raw = request.query_params.get("detail")
142-170: Tighten typing and avoid brittle ChatRequest detection.
- Add return type.
- String‑matching “ChatRequest” can false‑trigger. Prefer a type/marker check if available (e.g., issubclass(schema, ChatRequest) or hasattr(schema, "nat_chat_request")).
Apply this diff:
- def build_schema_info(fn: Function): + def build_schema_info(fn: Function) -> dict[str, Any] | None: schema = getattr(fn, "input_schema", None) is_chat_request = False if schema is not None: schema_name_part = getattr(schema, "__name__", None) qual_name_part = getattr(schema, "__qualname__", "") schema_name = schema_name_part or qual_name_part if schema_name and "ChatRequest" in str(schema_name): is_chat_request = TrueFollow‑up: if a concrete ChatRequest type is available in this module, switch to an
issubclasscheck.
188-193: Use precise workflow detection.
hasattr(fn, "run")may produce false positives. Preferisinstance(fn, Workflow)(already imported).Apply this diff:
selected_entry: dict[str, Any] = { "name": n, "description": get_function_description(fn), - "is_workflow": hasattr(fn, "run"), + "is_workflow": isinstance(fn, Workflow), }
209-222: Always include is_workflow for consistent shape.Clients shouldn’t have to branch on
detailto know if an entry is a workflow.Apply this diff:
for name, fn in functions.items(): list_entry: dict[str, Any] = { "name": name, "description": get_function_description(fn), + "is_workflow": isinstance(fn, Workflow), } - if include_schemas: - list_entry["schema"] = build_schema_info(fn) - list_entry["is_workflow"] = hasattr(fn, "run") + if include_schemas: + list_entry["schema"] = build_schema_info(fn) tools.append(list_entry)
113-208: Add basic route tests (Starlette TestClient).Recommend adding coverage for: no params, names (single/multi), detail toggles, 404 on missing, schema presence for ChatRequest.
I can generate a minimal pytest using Starlette’s TestClient around FastMCP to assert response shapes. Want me to open a follow‑up PR with tests?
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
docs/source/workflows/mcp/mcp-server.md(1 hunks)src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py(3 hunks)
🧰 Additional context used
📓 Path-based instructions (3)
**/*
⚙️ CodeRabbit configuration file
**/*: # Code Review Instructions
- Ensure the code follows best practices and coding standards. - For Python code, follow
PEP 20 and
PEP 8 for style guidelines.- Check for security vulnerabilities and potential issues. - Python methods should use type hints for all parameters and return values.
Example:def my_function(param1: int, param2: str) -> bool: pass- For Python exception handling, ensure proper stack trace preservation:
- When re-raising exceptions: use bare
raisestatements to maintain the original stack trace,
and uselogger.error()(notlogger.exception()) to avoid duplicate stack trace output.- When catching and logging exceptions without re-raising: always use
logger.exception()
to capture the full stack trace information.Documentation Review Instructions - Verify that documentation and comments are clear and comprehensive. - Verify that the documentation doesn't contain any TODOs, FIXMEs or placeholder text like "lorem ipsum". - Verify that the documentation doesn't contain any offensive or outdated terms. - Verify that documentation and comments are free of spelling mistakes, ensure the documentation doesn't contain any
words listed in the
ci/vale/styles/config/vocabularies/nat/reject.txtfile, words that might appear to be
spelling mistakes but are listed in theci/vale/styles/config/vocabularies/nat/accept.txtfile are OK.Misc. - All code (except .mdc files that contain Cursor rules) should be licensed under the Apache License 2.0,
and should contain an Apache License 2.0 header comment at the top of each file.
- Confirm that copyright years are up-to date whenever a file is changed.
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.pydocs/source/workflows/mcp/mcp-server.md
src/nat/**/*
⚙️ CodeRabbit configuration file
This directory contains the core functionality of the toolkit. Changes should prioritize backward compatibility.
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
docs/source/**/*
⚙️ CodeRabbit configuration file
This directory contains the source code for the documentation. All documentation should be written in Markdown format. Any image files should be placed in the
docs/source/_staticdirectory.
Files:
docs/source/workflows/mcp/mcp-server.md
Signed-off-by: Daniel Wang <daniewang@nvidia.com>
Signed-off-by: Daniel Wang <daniewang@nvidia.com>
There was a problem hiding this comment.
Actionable comments posted: 1
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py (1)
243-245: Gate debug endpoints behind config/flag.These routes enumerate internal tools/schemas; expose only when explicitly enabled (e.g., config flag or env var) or add simple auth.
Example:
+import os @@ - self._setup_debug_endpoints(mcp, functions) + if os.getenv("NAT_MCP_DEBUG_ROUTES", "0") == "1": + self._setup_debug_endpoints(mcp, functions)Or add
debug_routes_enabled: bool = FalsetoMCPFrontEndConfigand check it here.
🧹 Nitpick comments (6)
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py (3)
124-131: Preserve request order and deduplicate deterministically.Using
set(...)loses order and can yield unstable output. Build an ordered, de‑duplicated list.- names_param_list = set(request.query_params.getlist("name")) - names: list[str] = [] - for raw in names_param_list: - # if p.strip() is empty, it won't be included in the list! - parts = [p.strip() for p in raw.split(",") if p.strip()] - names.extend(parts) + raw_values = request.query_params.getlist("name") + names: list[str] = [] + seen: set[str] = set() + for raw in raw_values: + for p in (s.strip() for s in raw.split(",")): + if not p or p in seen: + continue + seen.add(p) + names.append(p)
189-195: Consistent JSON error shape for 404s.Raising
HTTPExceptionyields Starlette’s default body; consider returning a JSON shape consistent with success responses (helps clients).- except KeyError as e: - raise HTTPException(status_code=404, detail=f"Tool \"{e.args[0]}\" not found.") from e + except KeyError as e: + from starlette.responses import JSONResponse + return JSONResponse( + {"count": 0, "tools": [], "server_name": mcp.name, "error": f'Tool "{e.args[0]}" not found.'}, + status_code=404)
199-201: Avoid caching debug responses.Add
Cache-Control: no-storeto prevent intermediaries from caching potentially sensitive metadata.- return JSONResponse( - _build_final_json(functions_to_include, _parse_detail_param(detail_raw, has_names=bool(names)))) + return JSONResponse( + _build_final_json(functions_to_include, _parse_detail_param(detail_raw, has_names=bool(names))), + headers={"Cache-Control": "no-store"})tests/nat/front_ends/mcp/test_mcp_debug_routes.py (3)
35-39: Avoid mutable default for lists in Pydantic models.Use
Field(default_factory=list).-from pydantic import BaseModel +from pydantic import BaseModel, Field @@ -class ChatRequestSchema(BaseModel): +class ChatRequestSchema(BaseModel): """Mock ChatRequest schema for testing special handling.""" - messages: list = [] + messages: list = Field(default_factory=list) model: str | None = None
126-131: Tests rely on a private method.Prefer going through the public
add_routes(...)to reduce coupling, or clearly mark this as intentional unit‑level access.If switching is feasible, confirm whether a lightweight
WorkflowBuilderstub exists to registertest_functionswithout full workflow wiring.
137-141: Drop asyncio markers for sync TestClient tests.These tests are sync;
@pytest.mark.asyncioandasync defadd overhead.Example:
- @pytest.mark.asyncio - async def test_route_exists_and_structure(...): + def test_route_exists_and_structure(...): ...Apply similarly to other tests in this module.
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py(3 hunks)tests/nat/front_ends/mcp/test_mcp_debug_routes.py(1 hunks)
🧰 Additional context used
📓 Path-based instructions (9)
src/**/*.py
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
src/**/*.py: All importable Python code must live under src/
All public APIs in src/ require Python 3.11+ type hints on parameters and return values; prefer typing/collections.abc abstractions; use typing.Annotated when useful
Provide Google-style docstrings for every public module, class, function, and CLI command; first line concise with a period; surround code entities with backticks
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
src/nat/**/*
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
Core functionality under src/nat should prioritize backward compatibility when changed
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
⚙️ CodeRabbit configuration file
This directory contains the core functionality of the toolkit. Changes should prioritize backward compatibility.
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
**/*.py
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
**/*.py: Follow PEP 8/20 style; format with yapf (column_limit=120) and use 4-space indentation; end files with a single newline
Run ruff (ruff check --fix) per pyproject.toml; fix warnings unless explicitly ignored; ruff is linter-only
Use snake_case for functions/variables, PascalCase for classes, and UPPER_CASE for constants
Treat pyright warnings as errors during development
Exception handling: preserve stack traces and avoid duplicate logging
When re-raising exceptions, use bareraiseand log with logger.error(), not logger.exception()
When catching and not re-raising, log with logger.exception() to capture stack trace
Validate and sanitize all user input; prefer httpx with SSL verification and follow OWASP Top‑10
Use async/await for I/O-bound work; profile CPU-heavy paths with cProfile/mprof; cache with functools.lru_cache or external cache; leverage NumPy vectorization when beneficial
**/*.py: Programmatic use: create TestLLMConfig(response_seq=[...], delay_ms=...), add with builder.add_llm("", cfg).
When retrieving the test LLM wrapper, use builder.get_llm(name, wrapper_type=LLMFrameworkEnum.) and call the framework’s method (e.g., ainvoke, achat, call).
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.pytests/nat/front_ends/mcp/test_mcp_debug_routes.py
**/*.{py,sh,md,yml,yaml,toml,ini,json,ipynb,txt,rst}
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
**/*.{py,sh,md,yml,yaml,toml,ini,json,ipynb,txt,rst}: Every file must start with the standard SPDX Apache-2.0 header; keep copyright years up‑to‑date
All source files must include the SPDX Apache‑2.0 header; do not bypass CI header checks
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.pytests/nat/front_ends/mcp/test_mcp_debug_routes.py
**/*.{py,md}
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
Never hard‑code version numbers in code or docs; versions are derived by setuptools‑scm
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.pytests/nat/front_ends/mcp/test_mcp_debug_routes.py
**/*.{py,yaml,yml}
📄 CodeRabbit inference engine (.cursor/rules/nat-test-llm.mdc)
**/*.{py,yaml,yml}: Configure response_seq as a list of strings; values cycle per call, and [] yields an empty string.
Configure delay_ms to inject per-call artificial latency in milliseconds for nat_test_llm.
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.pytests/nat/front_ends/mcp/test_mcp_debug_routes.py
**/*
⚙️ CodeRabbit configuration file
**/*: # Code Review Instructions
- Ensure the code follows best practices and coding standards. - For Python code, follow
PEP 20 and
PEP 8 for style guidelines.- Check for security vulnerabilities and potential issues. - Python methods should use type hints for all parameters and return values.
Example:def my_function(param1: int, param2: str) -> bool: pass- For Python exception handling, ensure proper stack trace preservation:
- When re-raising exceptions: use bare
raisestatements to maintain the original stack trace,
and uselogger.error()(notlogger.exception()) to avoid duplicate stack trace output.- When catching and logging exceptions without re-raising: always use
logger.exception()
to capture the full stack trace information.Documentation Review Instructions - Verify that documentation and comments are clear and comprehensive. - Verify that the documentation doesn't contain any TODOs, FIXMEs or placeholder text like "lorem ipsum". - Verify that the documentation doesn't contain any offensive or outdated terms. - Verify that documentation and comments are free of spelling mistakes, ensure the documentation doesn't contain any
words listed in the
ci/vale/styles/config/vocabularies/nat/reject.txtfile, words that might appear to be
spelling mistakes but are listed in theci/vale/styles/config/vocabularies/nat/accept.txtfile are OK.Misc. - All code (except .mdc files that contain Cursor rules) should be licensed under the Apache License 2.0,
and should contain an Apache License 2.0 header comment at the top of each file.
- Confirm that copyright years are up-to date whenever a file is changed.
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.pytests/nat/front_ends/mcp/test_mcp_debug_routes.py
tests/**/*.py
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
Unit tests must live under tests/ and use configured markers (e2e, integration, etc.)
Files:
tests/nat/front_ends/mcp/test_mcp_debug_routes.py
⚙️ CodeRabbit configuration file
tests/**/*.py: - Ensure that tests are comprehensive, cover edge cases, and validate the functionality of the code. - Test functions should be named using thetest_prefix, using snake_case. - Any frequently repeated code should be extracted into pytest fixtures. - Pytest fixtures should define the name argument when applying the pytest.fixture decorator. The fixture
function being decorated should be named using thefixture_prefix, using snake_case. Example:
@pytest.fixture(name="my_fixture")
def fixture_my_fixture():
pass
Files:
tests/nat/front_ends/mcp/test_mcp_debug_routes.py
**/tests/**/*.py
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
**/tests/**/*.py: Test functions must use the test_ prefix and snake_case
Extract repeated test code into pytest fixtures; fixtures should set name=... in @pytest.fixture and functions named with fixture_ prefix
Mark expensive tests with @pytest.mark.slow or @pytest.mark.integration
Use pytest with pytest-asyncio for async code; mock external services with pytest_httpserver or unittest.mock
Files:
tests/nat/front_ends/mcp/test_mcp_debug_routes.py
🧬 Code graph analysis (2)
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py (3)
src/nat/builder/function.py (1)
Function(46-301)src/nat/front_ends/mcp/tool_converter.py (1)
get_function_description(199-236)src/nat/data_models/common.py (1)
model_json_schema(135-151)
tests/nat/front_ends/mcp/test_mcp_debug_routes.py (2)
src/nat/data_models/config.py (1)
GeneralConfig(188-240)src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py (1)
_setup_debug_endpoints(107-200)
Signed-off-by: Daniel Wang <daniewang@nvidia.com>
There was a problem hiding this comment.
Actionable comments posted: 0
♻️ Duplicate comments (1)
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py (1)
89-107: Fix return type and variable annotation: mapping contains Workflow, not just Function.
functionsincludesworkflow(aWorkflow) at Line 105, sodict[str, Function]is incorrect and will trip pyright. UseFunctionBasein both the return type and local var; update docstring accordingly.Apply:
-def _get_all_functions(self, workflow: Workflow) -> dict[str, Function]: - """Get all functions from the workflow. +def _get_all_functions(self, workflow: Workflow) -> dict[str, FunctionBase]: + """Get all functions and the workflow entry. @@ - Returns: - Dict mapping function names to Function objects. + Returns: + Dict mapping tool names to FunctionBase (Function or Workflow) instances. @@ - functions: dict[str, Function] = {} + functions: dict[str, FunctionBase] = {} @@ - functions[workflow.config.workflow.type] = workflow + functions[workflow.config.workflow.type] = workflow
🧹 Nitpick comments (5)
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py (5)
109-116: Docstring nit: hyphenate “comma‑separated”.Minor clarity tweak.
Apply:
- - GET /debug/tools/list: List tools. Optional query param `name` (one or more, repeatable or comma separated) + - GET /debug/tools/list: List tools. Optional query param `name` (one or more, repeatable or comma‑separated)
127-133: Dedup names while preserving caller order (avoid set reordering).Using
set()first can reorder inputs; build an ordered, deduped list instead.Apply:
-names_param_list = set(request.query_params.getlist("name")) -names: list[str] = [] -for raw in names_param_list: - # if p.strip() is empty, it won't be included in the list! - parts = [p.strip() for p in raw.split(",") if p.strip()] - names.extend(parts) +raw_values = request.query_params.getlist("name") +names: list[str] = [] +seen: set[str] = set() +for raw in raw_values: + for p in raw.split(","): + name = p.strip() + if name and name not in seen: + seen.add(name) + names.append(name)
149-175: Make schema detection more robust; add Pydantic v1 fallback.
- Prefer an
issubclasscheck for ChatRequest when importable; keep name heuristic as fallback.- Support Pydantic v1 (
schema()) in addition to v2 (model_json_schema()).Apply:
def _build_schema_info(fn: FunctionBase) -> dict[str, Any] | None: schema = getattr(fn, "input_schema", None) if schema is None: return None - # check if schema is a ChatRequest + # Try explicit ChatRequest type check; fall back to name heuristic + try: + from nat.front_ends.mcp.data_models import ChatRequest as _ChatRequest # type: ignore + except Exception: + _ChatRequest = None # type: ignore[assignment] + schema_name = getattr(schema, "__name__", "") schema_qualname = getattr(schema, "__qualname__", "") - if "ChatRequest" in schema_name or "ChatRequest" in schema_qualname: + if (_ChatRequest and isinstance(schema, type) and issubclass(schema, _ChatRequest)) or \ + ("ChatRequest" in schema_name or "ChatRequest" in schema_qualname): # Simplified interface used by MCP wrapper for ChatRequest return { "type": "object", "properties": { "query": { "type": "string", "description": "User query string" } }, "required": ["query"], "title": "ChatRequestQuery", } - # Pydantic models provide model_json_schema - if schema is not None and hasattr(schema, "model_json_schema"): - return schema.model_json_schema() + # Pydantic models: v2 (model_json_schema) and v1 (schema) + if schema is not None: + if hasattr(schema, "model_json_schema"): + return schema.model_json_schema() + if hasattr(schema, "schema"): + return schema.schema() return None
176-192: Use isinstance for workflow detection and wrap to stay within 120 cols.More precise and keeps formatting within style limits.
Apply:
- for name, fn in functions_to_include.items(): - list_entry: dict[str, Any] = { - "name": name, "description": get_function_description(fn), "is_workflow": hasattr(fn, "run") - } + for name, fn in functions_to_include.items(): + list_entry: dict[str, Any] = { + "name": name, + "description": get_function_description(fn), + "is_workflow": isinstance(fn, Workflow), + }
247-248: Gate debug endpoints behind a config flag to avoid accidental exposure.To reduce info‑disclosure risk in prod, let users disable this route via config. Default to enabled for backward compatibility.
Apply:
-# After registration, expose debug endpoints for tool/schema inspection -self._setup_debug_endpoints(mcp, functions) +# After registration, optionally expose debug endpoints for tool/schema inspection +if getattr(self.front_end_config, "enable_debug_routes", True): + self._setup_debug_endpoints(mcp, functions) +else: + logger.info("MCP debug endpoints disabled via front_end_config.enable_debug_routes=False")
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py(3 hunks)
🧰 Additional context used
📓 Path-based instructions (7)
src/**/*.py
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
src/**/*.py: All importable Python code must live under src/
All public APIs in src/ require Python 3.11+ type hints on parameters and return values; prefer typing/collections.abc abstractions; use typing.Annotated when useful
Provide Google-style docstrings for every public module, class, function, and CLI command; first line concise with a period; surround code entities with backticks
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
src/nat/**/*
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
Core functionality under src/nat should prioritize backward compatibility when changed
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
⚙️ CodeRabbit configuration file
This directory contains the core functionality of the toolkit. Changes should prioritize backward compatibility.
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
**/*.py
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
**/*.py: Follow PEP 8/20 style; format with yapf (column_limit=120) and use 4-space indentation; end files with a single newline
Run ruff (ruff check --fix) per pyproject.toml; fix warnings unless explicitly ignored; ruff is linter-only
Use snake_case for functions/variables, PascalCase for classes, and UPPER_CASE for constants
Treat pyright warnings as errors during development
Exception handling: preserve stack traces and avoid duplicate logging
When re-raising exceptions, use bareraiseand log with logger.error(), not logger.exception()
When catching and not re-raising, log with logger.exception() to capture stack trace
Validate and sanitize all user input; prefer httpx with SSL verification and follow OWASP Top‑10
Use async/await for I/O-bound work; profile CPU-heavy paths with cProfile/mprof; cache with functools.lru_cache or external cache; leverage NumPy vectorization when beneficial
**/*.py: Programmatic use: create TestLLMConfig(response_seq=[...], delay_ms=...), add with builder.add_llm("", cfg).
When retrieving the test LLM wrapper, use builder.get_llm(name, wrapper_type=LLMFrameworkEnum.) and call the framework’s method (e.g., ainvoke, achat, call).
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
**/*.{py,sh,md,yml,yaml,toml,ini,json,ipynb,txt,rst}
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
**/*.{py,sh,md,yml,yaml,toml,ini,json,ipynb,txt,rst}: Every file must start with the standard SPDX Apache-2.0 header; keep copyright years up‑to‑date
All source files must include the SPDX Apache‑2.0 header; do not bypass CI header checks
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
**/*.{py,md}
📄 CodeRabbit inference engine (.cursor/rules/general.mdc)
Never hard‑code version numbers in code or docs; versions are derived by setuptools‑scm
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
**/*.{py,yaml,yml}
📄 CodeRabbit inference engine (.cursor/rules/nat-test-llm.mdc)
**/*.{py,yaml,yml}: Configure response_seq as a list of strings; values cycle per call, and [] yields an empty string.
Configure delay_ms to inject per-call artificial latency in milliseconds for nat_test_llm.
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
**/*
⚙️ CodeRabbit configuration file
**/*: # Code Review Instructions
- Ensure the code follows best practices and coding standards. - For Python code, follow
PEP 20 and
PEP 8 for style guidelines.- Check for security vulnerabilities and potential issues. - Python methods should use type hints for all parameters and return values.
Example:def my_function(param1: int, param2: str) -> bool: pass- For Python exception handling, ensure proper stack trace preservation:
- When re-raising exceptions: use bare
raisestatements to maintain the original stack trace,
and uselogger.error()(notlogger.exception()) to avoid duplicate stack trace output.- When catching and logging exceptions without re-raising: always use
logger.exception()
to capture the full stack trace information.Documentation Review Instructions - Verify that documentation and comments are clear and comprehensive. - Verify that the documentation doesn't contain any TODOs, FIXMEs or placeholder text like "lorem ipsum". - Verify that the documentation doesn't contain any offensive or outdated terms. - Verify that documentation and comments are free of spelling mistakes, ensure the documentation doesn't contain any
words listed in the
ci/vale/styles/config/vocabularies/nat/reject.txtfile, words that might appear to be
spelling mistakes but are listed in theci/vale/styles/config/vocabularies/nat/accept.txtfile are OK.Misc. - All code (except .mdc files that contain Cursor rules) should be licensed under the Apache License 2.0,
and should contain an Apache License 2.0 header comment at the top of each file.
- Confirm that copyright years are up-to date whenever a file is changed.
Files:
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py
🧬 Code graph analysis (1)
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py (2)
src/nat/builder/function.py (1)
Function(46-301)src/nat/front_ends/mcp/tool_converter.py (1)
get_function_description(199-236)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: CI Pipeline / Check
🔇 Additional comments (3)
src/nat/front_ends/mcp/mcp_front_end_plugin_worker.py (3)
193-201: LGTM: 404 on first missing tool with clear message.Behavior is sensible for a debug endpoint.
203-205: LGTM: defaultingdetailbased on presence ofnameis intuitive.Nice UX touch; response envelope looks consistent with
/health.
19-21: Imports look right for new types and exceptions.Good use of
Mapping/FunctionBaseand Starlette’sHTTPException.Also applies to: 23-23, 27-27
zhongxuanwang-nv
changed the title
3 participants
Description
Closes AIQ-1851
This will be a custom MCP route similar to "/health" and used for displaying tools and their schema without the need for an MCP client.
By Submitting this PR I confirm:
Summary by CodeRabbit
New Features
Documentation
Tests