Skip to content
docs

CrewAI and Deedglass

Give a CrewAI agent Deedglass’s tools in one file. The file starts Deedglass’s MCP server (npx -y deedglass mcp serve) and hands its tools to the agent: each call goes through your policy, the risky ones wait for a person, and each one is sealed as a deed you can check later.

Tested with crewai and crewai-tools[mcp] 1.15.28 (2026-10-21; CrewAI needs Python below 3.14).

Before you start

Section titled “Before you start”
  1. In any folder, run npx deedglass init (Docker, about 15 seconds; the quickstart). It starts Deedglass and a demo Postgres with 500 orders, and saves an agent key in .deedglass/quickstart.key.
  2. Run the file below from that folder (or a folder inside it). deedglass mcp serve looks for the key there; from anywhere else, set DEEDGLASS_API_KEY and DEEDGLASS_SERVER yourself (how mcp serve finds them).
  3. Under the postgres-safe policy that init turns on, reads run, DROP and TRUNCATE are refused, and a write waits for a person in Console › Approvals before it runs.

The MCP route

Section titled “The MCP route”
examples/frameworks/crewai/agent.py
# Run where you ran `npx deedglass init`. Needs OPENAI_API_KEY (CrewAI's default model).
from crewai import Agent, Crew, Task
from crewai_tools import MCPServerAdapter
from mcp import StdioServerParameters


params = StdioServerParameters(command="npx", args=["-y", "deedglass", "mcp", "serve"])
with MCPServerAdapter(params, connect_timeout=60) as mcp_tools:  # CrewAI names them postgres_query, ...
    analyst = Agent(role="Orders analyst", goal="Answer questions about orders from the database.",
                    backstory="You read the orders database through Deedglass.", tools=mcp_tools)
    task = Task(description="How many rows are in public.orders?", expected_output="One number.", agent=analyst)
    print(Crew(agents=[analyst], tasks=[task]).kickoff())

Install and run:

Terminal window
pip install crewai==1.15.28 "crewai-tools[mcp]==1.15.28"
python agent.py

Model key: OPENAI_API_KEY, for CrewAI’s default model.

Tool names: CrewAI’s MCP adapter lists the tools as postgres_query and so on, as Deedglass names them. Policies and records keep the dotted name, postgres.query (MCP.md).

What you see: the answer (500 rows), and in Console › Ledger or npx deedglass actions list one sealed postgres.query action per call, with the rule that allowed it. If the agent tries a write, the tool answers that it is waiting for approval and nothing has run; once someone approves, deedglass_check_approval lets the agent carry on. A refused call answers with the policy’s sentence and a record id.

First start: npx -y deedglass can take several seconds the first time. If a framework’s MCP connect timeout is shorter, run npx -y deedglass version once first, or install the CLI (INSTALL.md) and use deedglass as the command.

Check it without a model key

Section titled “Check it without a model key”

examples/frameworks/crewai/ has the file, its pinned packages and a check file that needs no model key: it starts deedglass mcp serve through the same framework client, lists its tools and gives them to an Agent. From the folder where init ran:

Terminal window
path/to/deedglass/examples/frameworks/run.sh crewai            # install the pins in a temp folder, run the check
path/to/deedglass/examples/frameworks/run.sh crewai --agent    # also run the agent file when its model key is set

The workflow frameworks.yml runs every check against a fresh init stack each week and on pull requests that touch these files. It isn’t a required check: a framework release that doesn’t install is reported as skipped, not as a failure.

Sources and what isn’t verified

Section titled “Sources and what isn’t verified”

Every framework API in the files above comes from the framework’s own docs, checked 2026-10-21:

Not verified:

  • Task(description, expected_output, agent), Crew(agents, tasks).kickoff(): CrewAI’s MCP page stops at the Agent, so these come from CrewAI’s core classes as installed.
  • A run with a live model (no model key in CI).