Loading ToolBench

Reading the current corpus…

zhivex-ai-sdk-py MCP server · F · ToolBench

← Browse

zhivex-ai-sdk-py

github.com/Zhivex/zhivex-ai-sdk-py

Provider-portable Python runtime for building reliable AI agents

2026-07-28+STDIO8 tools

36/100

Definition Quality42

Protocol Readiness0

Supportability50

Spec posture

Inferred effective spec 2026-07-28+·current baseline 2026-07-28

Static source inference · medium confidence

Current-spec patterns detected

Summary

This is not an MCP server, it is a Python AI SDK (Zhivex) with example agent scripts that happen to reference tools. The 8 'tools' are scattered across example files (messages_and_tools.py, remote_tool.py, dev_agent_gemini_search_tool.py, etc.) and are NOT formally registered as MCP tools. Tool definitions are inferred from docstrings and function signatures, not explicit MCP tool registration. No MCP server code is present in the repository. The tools lack consistent schemas, parameter descriptions, and error handling.

Tools (8)

delegate_researchread onlysource verified53/100

Delegates the current task to the researcher agent.

lookupread onlysource verified27/100

Returns the marker required to answer the user.

project_statusread onlyauthsource verified37/100

Fetches project status from a remote HTTP tool endpoint.

pythonwriteauthsource verified28/100

Vertex Gemini Python code execution tool

search_webread onlysource verified57/100

Runs a grounded web search with Gemini and returns an answer plus sources.

urlread onlyauthsource verified30/100

Vertex URL context retrieval tool

weatherread onlysource verified

Returns a tiny forecast snapshot for a city.

weather

Issues (5)

CRITICALtool ↗

Not an MCP server. This is a Python SDK with example scripts. No formal MCP tool registration found. Tools are inferred from scattered example files, not explicitly registered via MCP protocol.

weatherproject_statussearch_weblookupdelegate_researchpythonurl

HIGHtool ↗

Duplicate tool names. Two 'weather' tools with identical signatures but different descriptions ('tiny weather summary' vs 'tiny forecast snapshot'). LLMs cannot distinguish between them.

weather

CRITICALconstrained-input ↗

Missing input schemas for 5 tools. project_status, lookup, python, and url have no visible parameter schemas in the source.

project_statuslookuppythonurl

HIGHtool ↗

Vague tool names. 'lookup' and 'python' do not start with action verbs and do not convey what they do. 'lookup' could mean search, retrieve, or resolve. 'python' is generic, does it execute, compile, or analyze?

Recommendations

Convert this repository into a formal MCP server using the mcp Python library (already in optional dependencies). Register all tools via the MCP protocol with explicit tool definitions, not scattered example scripts.
Rename duplicate 'weather' tools to 'get_weather_current' and 'get_weather_forecast' to disambiguate. Add descriptions explaining the difference (current conditions vs multi-day forecast).
Rename 'lookup' to 'lookup_marker' or 'get_marker' and 'python' to 'execute_python' or 'run_python_code'. Start all tool names with action verbs.
Add complete input schemas for all tools. For tools with no parameters (project_status, lookup), explicitly declare 'properties: {}' and 'additionalProperties: false'. For python and url, document all accepted parameters with types and descriptions.
Expand descriptions to 50 - 200 characters. Include WHAT the tool does, WHEN to use it, and WHAT it returns. Example: 'lookup' → 'Retrieves the marker ID required to answer the user query. Call this first to identify the correct marker before proceeding.'

Score history

Overall score trend

First recorded score · v2 rubric

36/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	36	2026-07-28+	v2