agentcli generate

Generate deployment artifacts from an agent blueprint without deploying. Produces the .bedrock_agentcore.yaml runtime configuration and a Dockerfile that agentcli deploy also generates internally.

Use generate when you want to inspect or customize artifacts before deployment, or when you manage your own CI/CD pipeline and want to separate artifact generation from the deploy step.

Synopsis 

agentcli generate <subcommand> [OPTIONS] YAML_PATH

Subcommands

Subcommand Description
runtime-config Generate .bedrock_agentcore.yaml from an agent blueprint
dockerfile Generate a Dockerfile from an agent blueprint

agentcli generate runtime-config

Generate the .bedrock_agentcore.yaml configuration file that the AgentCore starter toolkit reads to configure the runtime.

agentcli generate runtime-config YAML_PATH [OPTIONS]

Arguments

Argument Description
YAML_PATH Path to the agent blueprint YAML file

Options

Option Short Default Description
--output-dir -o . Directory to write .bedrock_agentcore.yaml into
--entrypoint -e app.py Python entrypoint file name

Example 

# Generate into current directory
agentcli generate runtime-config agents/my-agent.yaml
# Generated: .bedrock_agentcore.yaml

# Generate into a specific directory
agentcli generate runtime-config agents/my-agent.yaml \
  --output-dir build/ \
  --entrypoint handler.py
# Generated: build/.bedrock_agentcore.yaml

agentcli generate dockerfile

Generate a production Dockerfile from an agent blueprint.

agentcli generate dockerfile YAML_PATH [OPTIONS]

Arguments

Argument Description
YAML_PATH Path to the agent blueprint YAML file

Options

Option Short Default Description
--output-dir -o . Directory to write Dockerfile into
--base-image   python:3.12-slim Docker base image
--requirements -r requirements.txt Requirements file to install

Examples 

# Generate with defaults
agentcli generate dockerfile agents/my-agent.yaml
# Generated: ./Dockerfile

# Specify a custom base image and requirements file
agentcli generate dockerfile agents/my-agent.yaml \
  --base-image python:3.12-slim \
  --requirements requirements-prod.txt \
  --output-dir build/
# Generated: build/Dockerfile

Generated Dockerfile Format

The generated Dockerfile follows the AgentCore container contract: expose port 8080, run via agent_core.runtime.entrypoint.

FROM python:3.12-slim

WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

COPY . .

CMD ["python", "-m", "agent_core.runtime.entrypoint"]

The agent_core.runtime.entrypoint module starts an HTTP server on port 8080 and registers your blueprint’s entrypoint handler for the /invocations route.

CI/CD Usage

In a CI/CD pipeline, generate and inspect artifacts before the deploy step:

# Step 1: Validate the blueprint
agentcli blueprint lint agents/my-agent.yaml

# Step 2: Generate artifacts
agentcli generate runtime-config agents/my-agent.yaml --output-dir dist/
agentcli generate dockerfile agents/my-agent.yaml --output-dir dist/

# Step 3: Inspect or commit the artifacts
cat dist/.bedrock_agentcore.yaml
cat dist/Dockerfile

# Step 4: Build and push the container image separately
docker build -f dist/Dockerfile -t my-agent:$VERSION .
docker push $ECR_REPO/my-agent:$VERSION

Difference from agentcli deploy

  agentcli generate agentcli deploy
Generates .bedrock_agentcore.yaml Yes Yes
Generates Dockerfile Yes Yes
Builds Docker image No Yes
Pushes to ECR No Yes
Creates AgentCore Runtime No Yes

Use generate for inspection, customization, or custom CI pipelines. Use deploy for a one-command full deployment.

See Also