Menü
Add a new LangChain tool to the Medox ReAct agent. Use when creating a new tool file, registering it in the graph, and writing its test.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
Run the LangSmith evaluation suite and display pass/fail results
Create a git commit with Medox conventions (co-author, conventional type/scope). Invoke manually with /commit — do NOT trigger automatically.
Audit code changes on the current branch and update docs/ + .claude/rules/ to stay in sync. Run before /pr.
Create or update a GitHub pull request following Medox conventions (author spicode-bot, reviewer spideystreet). Invoke manually with /pr — do NOT trigger automatically.
Fetch and display the full status of the current PR — CI checks, GitHub Actions runs, reviews, and code comments — using GitHub MCP tools. Invoke manually with /pr-status.
Rebuild ChromaDB Gold layer indexes after source data changes. Invoke manually — deletes and recreates collections, do NOT trigger automatically.
| name | add-agent-tool |
| description | Add a new LangChain tool to the Medox ReAct agent. Use when creating a new tool file, registering it in the graph, and writing its test. |
| argument-hint | <tool-name> |
Create the tool file at src/medox/agent/tools/tool_<name>.py:
"""One-line description of the data source and lookup strategy."""
from langchain_core.tools import tool
from medox.pipeline.config_pipeline import PipelineSettings
@tool
def <tool_name>(<param>: str) -> str:
"""
<What this tool does — written for the LLM, not a human developer.>
Include:
- What the parameter means and expected format
- Example values or class names the LLM should pass
- What the tool returns and how to interpret it
Example: pass 'ANTIVITAMINES K' for warfarine, 'STATINES' for simvastatine.
"""
settings = PipelineSettings()
# ... implementation
Register the tool in src/medox/agent/graph_agent.py:
tools = [...] list passed to build_agent()Docstring rules (critical — the LLM uses this to decide when/how to call the tool):
Write the test at tests/agent/test_tool_<name>.py:
def test_<tool_name>_returns_expected():
result = <tool_name>.invoke({"<param>": "<known_value>"})
assert "<expected_substring>" in result
def test_<tool_name>_handles_unknown():
result = <tool_name>.invoke({"<param>": "nonexistent_xyz"})
assert result # returns graceful message, not exception
Update .claude/rules/agent.md — add the tool to the tools table.
Validate:
uv run pytest tests/agent/test_tool_<name>.py -v
uv run pytest tests/agent/ -v
tool_<name>.py — matches the @tool function name@tool per file