CrewAI Tools Integration: Let Agents Search, Read, and Actually Do Work

An agent without tools is like an engineer with plenty of theory but no keyboard.
Can talk a lot, but cannot ship.

Common built-in tools

Frequently used CrewAI tools include:

SerperDevTool: web search
FileReadTool: read files
FileWriterTool: write files
ScrapeWebsiteTool: scrape webpage content

Attach tools to an agent

from crewai import Agent
from crewai_tools import SerperDevTool
 
search_tool = SerperDevTool()
 
researcher = Agent(
    role="Researcher",
    goal="Find trustworthy sources about the topic",
    backstory="You validate facts before writing.",
    tools=[search_tool],
    allow_delegation=False,
    verbose=True,
)

💡 You can reuse the same search_tool across agents. No need to recreate it every time (DRY).

Minimal custom tool template

from crewai.tools import BaseTool
from pydantic import BaseModel, Field
 
class QueryInput(BaseModel):
    keyword: str = Field(description="Search keyword")
 
class KeywordTool(BaseTool):
    name: str = "Keyword Tool"
    description: str = "Return a short explanation for a keyword."
    args_schema: type[BaseModel] = QueryInput
 
    def _run(self, keyword: str) -> str:
        return f"Result for: {keyword}"

Four tool design rules

Do one thing only: if it searches, do not also write files
Be explicit: description should clearly state purpose and input
Validate parameters: use Pydantic schema
Return stable format: easier for downstream tasks to parse

Common errors

Symptom	Cause	Fix
Tool never called	Task prompt does not explicitly require it	Add explicit "use tool first" instruction in task
Bad parameter values	args schema is unclear	Add `Field(description=...)`
Hard-to-use output	Free-form text too scattered	Return fixed sections or JSON string

Next step

After tools, the next focus is output consistency.
Next post shows how to lock format with Pydantic:
👉 Structured Output with Pydantic