agent-starter-pack 0.6.3__py3-none-any.whl → 0.7.0__py3-none-any.whl

llm.txt

@@ -0,0 +1,290 @@
+---
+**llm.txt** documents the "Agent Starter Pack" repository, providing a source of truth on its purpose, features, and usage.
+---
+
+### Section 1: Project Overview
+
+*   **Project Name:** Agent Starter Pack
+*   **Purpose:** Accelerate development of production-ready GenAI Agents on Google Cloud.
+*   **Tagline:** Production-Ready Agents on Google Cloud, faster.
+
+**The "Production Gap":**
+While prototyping GenAI agents is quick, production deployment often takes 3-9 months.
+
+**Key Challenges Addressed:**
+*   **Customization:** Business logic, data grounding, security/compliance.
+*   **Evaluation:** Metrics, quality assessment, test datasets.
+*   **Deployment:** Cloud infrastructure, CI/CD, UI integration.
+*   **Observability:** Performance tracking, user feedback.
+
+**Solution: Agent Starter Pack**
+Provides MLOps and infrastructure templates so developers focus on agent logic.
+
+*   **You Build:** Prompts, LLM interactions, business logic, agent orchestration.
+*   **We Provide:**
+    *   Deployment infrastructure, CI/CD, testing
+    *   Logging, monitoring
+    *   Evaluation tools
+    *   Data connections, UI playground
+    *   Security best practices
+
+Establishes production patterns from day one, saving setup time.
+
+---
+### Section 2: Creating a New Agent Project
+
+Start by creating a new agent project from a predefined template. This process supports both interactive and fully automated setup.
+
+**Prerequisites:**
+Before you begin, ensure you have `uv`/`uvx`, `gcloud` CLI, `terraform`, `git`, and `gh` CLI (for automated CI/CD setup) installed and authenticated.
+
+**Installing the `agent-starter-pack` CLI:**
+Choose one method to get the `agent-starter-pack` command:
+
+1.  **`uvx` (Recommended for Zero-Install/Automation):** Run directly without prior installation.
+    ```bash
+    uvx agent-starter-pack create ...
+    ```
+2.  **Virtual Environment (`pip` or `uv`):**
+    ```bash
+    pip install agent-starter-pack
+    ```
+3.  **Persistent CLI Install (`pipx` or `uv tool`):** Installs globally in an isolated environment.
+
+---
+### `agent-starter-pack create` Command
+
+Generates a new agent project directory based on a chosen template and configuration.
+
+**Usage:**
+```bash
+agent-starter-pack create PROJECT_NAME [OPTIONS]
+```
+
+**Arguments:**
+*   `PROJECT_NAME`: Name for your new project directory and base for GCP resource naming.
+
+**Key Options:**
+*   `-a, --agent`: Agent template (e.g., `adk_base`, `agentic_rag`).
+*   `-d, --deployment-target`: Target deployment environment (`cloud_run` or `agent_engine`).
+*   `-ds, --datastore`: For RAG agents, the datastore (`vertex_ai_search` or `vertex_ai_vector_search`).
+*   `-i, --include-data-ingestion`: Include data ingestion pipeline scaffolding.
+*   `--region`: GCP region (e.g., `us-central1`).
+*   `--auto-approve`: **Skips all interactive prompts (crucial for automation).**
+*   `--debug`: Enables debug logging during creation.
+
+**Automated Creation Example:**
+```bash
+uvx agent-starter-pack create my-automated-agent \
+  -a adk_base \
+  -d cloud_run \
+  --region us-central1 \
+  --auto-approve
+```
+
+---
+
+### Available Agent Templates
+
+Templates for the `create` command (via `-a` or `--agent`):
+
+| Agent Name             | Description                                  |
+| :--------------------- | :------------------------------------------- |
+| `adk_base`             | Base ReAct agent (ADK)                       |
+| `adk_gemini_fullstack` | Production-ready fullstack research agent    |
+| `agentic_rag`          | RAG agent for document retrieval & Q&A       |
+| `langgraph_base_react` | Base ReAct agent (LangGraph)                 |
+| `crewai_coding_crew`   | Multi-agent collaborative coding assistance  |
+| `live_api`             | Real-time multimodal RAG agent               |
+
+---
+
+### Including a Data Ingestion Pipeline (for RAG agents)
+
+For RAG agents needing custom document search, enabling this option automates loading, chunking, embedding documents with Vertex AI, and storing them in a vector database.
+
+**How to enable:**
+```bash
+uvx agent-starter-pack create my-rag-agent \
+  -a agentic_rag \
+  -d cloud_run \
+  -i \
+  -ds vertex_ai_search \
+  --auto-approve
+```
+**Post-creation:** Follow your new project's `data_ingestion/README.md` to deploy the necessary infrastructure.
+
+---
+### Section 3: Development & Automated Deployment Workflow
+---
+
+This section describes the end-to-end lifecycle of an agent, with emphasis on automation.
+
+
+### 1. Local Development & Iteration
+
+Once your project is created, navigate into its directory to begin development.
+
+**First, install dependencies (run once):**
+```bash
+make install
+```
+
+**Next, test your agent. The recommended method is to use a programmatic script.**
+
+#### Programmatic Testing (Recommended Workflow)
+
+This method allows for quick, automated validation of your agent's logic.
+
+1.  **Create a script:** In the project's root directory, create a Python script named `run_agent.py`.
+2.  **Invoke the agent:** In the script, write code to programmatically call your agent with sample input and `print()` the output for inspection.
+    *   **Guidance:** If you're unsure or no guidance exists, you can look at files in the `tests/` directory for examples of how to import and call the agent's main function.
+    *   **Important:** This script is for simple validation. **Assertions are not required**, and you should not create a formal `pytest` file.
+3.  **Run the test:** Execute your script from the terminal using `uv`.
+    ```bash
+    uv run python run_agent.py
+    ```
+You can keep the test file for future testing.
+
+#### Manual Testing with the UI Playground (Optional)
+
+If the user needs to interact with your agent manually in a chat interface for debugging:
+
+1.  Run the following command to start the local web UI:
+    ```bash
+    make playground
+    ```
+    This is useful for human-in-the-loop testing and features hot-reloading.
+
+### 2. Deploying to a Cloud Development Environment
+Before setting up full CI/CD, you can deploy to a personal cloud dev environment.
+
+1.  **Set Project:** `gcloud config set project YOUR_DEV_PROJECT_ID`
+2.  **Provision Resources:** `make setup-dev-env` (uses Terraform).
+3.  **Deploy Backend:** `make backend` (builds and deploys the agent).
+
+### 3. Automated Production-Ready Deployment with CI/CD
+For reliable deployments, the `setup-cicd` command streamlines the entire process. It creates a GitHub repo, connects it to Cloud Build, provisions staging/prod infrastructure, and configures deployment triggers.
+
+**Automated CI/CD Setup Example (Recommended):**
+```bash
+# Run from the project root. This command will guide you or can be automated with flags.
+uvx agent-starter-pack setup-cicd
+```
+
+**CI/CD Workflow Logic:**
+*   **On Pull Request:** CI pipeline runs tests.
+*   **On Merge to `main`:** CD pipeline deploys to staging.
+*   **Manual Approval:** A manual approval step triggers the production deployment.
+
+---
+### Section 4: Key Features & Customization
+---
+
+### Deploying with a User Interface (UI)
+*   **Unified Deployment (for Dev/Test):** The backend and frontend can be packaged and served from a single Cloud Run service, secured with Identity-Aware Proxy (IAP).
+*   **Deploying with UI:** `make backend IAP=true`
+*   **Access Control:** After deploying with IAP, grant users the `IAP-secured Web App User` role in IAM to give them access.
+
+### Monitoring & Observability
+*   **Technology:** Uses OpenTelemetry to emit events to Google Cloud Trace and Logging.
+*   **Custom Tracer:** A custom tracer in `app/utils/tracing.py` handles large payloads by linking to GCS, overcoming default service limits.
+*   **Infrastructure:** A Log Router to sink data to BigQuery is provisioned by Terraform.
+
+---
+### Section 5: CLI Reference for CI/CD Setup
+---
+
+### `agent-starter-pack setup-cicd`
+Automates the entire CI/CD setup for a project on GitHub.
+
+**Usage:**
+```bash
+uvx agent-starter-pack setup-cicd [OPTIONS]
+```
+**Prerequisites:** Run from the project root. `uv`, `gh` CLI, `gcloud` CLI, and `terraform` must be installed and authenticated. Requires `Owner` role on the GCP projects.
+
+**Key Automation Options:**
+*   `--staging-project`, `--prod-project`, `--cicd-project`, `--dev-project`
+*   `--region`, `--repository-name`, `--repository-owner`
+*   `--auto-approve`: **Crucial for automation.** Skips all interactive prompts.
+
+**Automated Example:**
+```bash
+uvx agent-starter-pack setup-cicd \
+  --staging-project t-starter-staging \
+  --prod-project t-starter-prod \
+  --repository-name my-adk-project-repo \
+  --repository-owner eliasecchig \
+  --auto-approve
+```
+
+---
+### Section 6: Operational Guidelines for Coding Agents
+
+These guidelines are essential for interacting with the Agent Starter Pack project effectively.
+
+---
+
+### Principle 1: Code Preservation & Isolation
+
+When executing code modifications using tools like `replace` or `write_file`, your paramount objective is surgical precision. You **must alter only the code segments directly targeted** by the user's request, while **strictly preserving all surrounding and unrelated code.**
+
+**Mandatory Pre-Execution Verification:**
+
+Before finalizing any `new_string` for a `replace` operation, meticulously verify the following:
+
+1.  **Target Identification:** Clearly define the exact lines or expressions to be changed, based *solely* on the user's explicit instructions.
+2.  **Preservation Check:** Compare your proposed `new_string` against the `old_string`. Ensure all code, configuration values (e.g., `model`, `version`, `api_key`), comments, and formatting *outside* the identified target remain identical and verbatim.
+
+**Example: Adhering to Preservation**
+
+*   **User Request:** "Change the agent's instruction to be a recipe suggester."
+*   **Original Code Snippet:**
+    ```python
+    root_agent = Agent(
+        name="root_agent",
+        model="gemini-2.5-flash",
+        instruction="You are a helpful AI assistant."
+    )
+    ```
+*   **Incorrect Modification (VIOLATION):**
+    ```python
+    root_agent = Agent(
+        name="recipe_suggester",
+        model="gemini-1.5-flash", # UNINTENDED MUTATION - model was not requested to change
+        instruction="You are a recipe suggester."
+    )
+    ```
+*   **Correct Modification (COMPLIANT):**
+    ```python
+    root_agent = Agent(
+        name="recipe_suggester", # OK, related to new purpose
+        model="gemini-2.5-flash", # MUST be preserved
+        instruction="You are a recipe suggester." # OK, the direct target
+    )
+    ```
+
+**Critical Error:** Failure to adhere to this preservation principle is a critical error. Always prioritize the integrity of existing, unchanged code over the convenience of rewriting entire blocks.
+
+---
+
+### Principle 2: Workflow & Execution Best Practices
+
+*   **Standard Workflow:**
+    The validated end-to-end process is: `create` → `test` → `setup-cicd` → push to deploy. Trust this high-level workflow as the default for developing and shipping agents.
+
+*   **Agent Testing:**
+    *   **Avoid `make playground`** unless specifically instructed; it is designed for human interaction. Focus on programmatic testing.
+
+*   **Model Selection:**
+    *   **When using Gemini, prefer the 2.5 model family** for optimal performance and capabilities: "gemini-2.5-pro" and "gemini-2.5-flash"
+
+*   **Running Python Commands:**
+    *   Always use `uv` to execute Python commands within this repository (e.g., `uv run run_agent.py`).
+    *   Ensure project dependencies are installed by running `make install` before executing scripts.
+    *   Consult the project's `Makefile` and `README.md` for other useful development commands.
+
+*   **Further Reading & Troubleshooting:**
+    *   For questions about specific frameworks (e.g., LangGraph) or Google Cloud products (e.g., Cloud Run), their official documentation and online resources are the best source of truth.
+    *   **When encountering persistent errors or if you're unsure how to proceed after initial troubleshooting, a targeted Google Search is strongly recommended.** It is often the fastest way to find relevant documentation, community discussions, or direct solutions to your problem.

src/base_template/.gitignore

@@ -194,4 +194,5 @@ my_env.tfvars
 .streamlit_chats
 .saved_chats
 .env
-.requirements.txt
+.requirements.txt
+GEMINI.md

src/base_template/GEMINI.md

@@ -0,0 +1,5 @@
+Coding Agent guidance:
+{%- if "adk" in cookiecutter.tags %}
+{{ cookiecutter.adk_cheatsheet }}
+{%- endif %}
+{{ cookiecutter.llm_txt }}

src/base_template/Makefile

@@ -1,10 +1,12 @@
+# Install dependencies using uv package manager
 install:
-	@command -v uv >/dev/null 2>&1 || { echo "uv is not installed. Installing uv..."; curl -LsSf https://astral.sh/uv/0.6.12/install.sh | sh; source ~/.bashrc; }
-	uv sync --dev{% if cookiecutter.agent_name != 'live_api' and "adk" not in cookiecutter.tags %} --extra streamlit{%- endif %} --extra jupyter --frozen{% if cookiecutter.agent_name == 'live_api' %} && npm --prefix frontend install{%- endif %}
+	@command -v uv >/dev/null 2>&1 || { echo "uv is not installed. Installing uv..."; curl -LsSf https://astral.sh/uv/0.6.12/install.sh | sh; source $HOME/.local/bin/env; }
+	uv sync --dev{% if cookiecutter.agent_name != 'live_api' and "adk" not in cookiecutter.tags %} --extra streamlit{%- endif %} --extra jupyter{% if cookiecutter.agent_name == 'live_api' %} && npm --prefix frontend install{%- endif %}
 {%- if cookiecutter.settings.get("commands", {}).get("override", {}).get("install") %} && {{cookiecutter.settings.get("commands", {}).get("override", {}).get("install")}}{%- endif %}
 {%- if cookiecutter.settings.get("commands", {}).get("extra", {}) %}
 {%- for cmd_name, cmd_value in cookiecutter.settings.get("commands", {}).get("extra", {}).items() %}
 
+# {{ cmd_value.get("description") }}
 {{ cmd_name }}:
 {%- if cmd_value is mapping %}
 {%- if cmd_value.command is mapping and cookiecutter.deployment_target in cmd_value.command %}
@@ -17,6 +19,7 @@ install:
 {%- endif %}
 {%- endfor %}{%- endif %}
 
+# Launch local dev playground
 playground:
 {%- if cookiecutter.settings.get("commands", {}).get("override", {}).get("playground") %}
 	{{cookiecutter.settings.get("commands", {}).get("override", {}).get("playground")}}
@@ -44,10 +47,14 @@ playground:
 {%- endif %}
 {%- endif %}
 
+# Deploy the agent remotely
+{%- if cookiecutter.deployment_target == 'cloud_run' %}
+# Usage: make backend [IAP=true] [PORT=8080] - Set IAP=true to enable Identity-Aware Proxy, PORT to specify container port
+{%- endif %}
 backend:
 {%- if cookiecutter.deployment_target == 'cloud_run' %}
 	PROJECT_ID=$$(gcloud config get-value project) && \
-	gcloud run deploy {{cookiecutter.project_name}} \
+	gcloud beta run deploy {{cookiecutter.project_name}} \
 		--source . \
 		--memory "4Gi" \
 		--project $$PROJECT_ID \
@@ -55,31 +62,37 @@ backend:
 		--no-allow-unauthenticated \
 		--labels "created-by=adk" \
 		--set-env-vars \
-		"COMMIT_SHA=$(shell git rev-parse HEAD){%- if cookiecutter.data_ingestion %}{%- if cookiecutter.datastore_type == "vertex_ai_search" %},DATA_STORE_ID={{cookiecutter.project_name}}-datastore,DATA_STORE_REGION=us{%- elif cookiecutter.datastore_type == "vertex_ai_vector_search" %},VECTOR_SEARCH_INDEX={{cookiecutter.project_name}}-vector-search,VECTOR_SEARCH_INDEX_ENDPOINT={{cookiecutter.project_name}}-vector-search-endpoint,VECTOR_SEARCH_BUCKET=$$PROJECT_ID-{{cookiecutter.project_name}}-vs{%- endif %}{%- endif %}"
+		"COMMIT_SHA=$(shell git rev-parse HEAD){%- if cookiecutter.data_ingestion %}{%- if cookiecutter.datastore_type == "vertex_ai_search" %},DATA_STORE_ID={{cookiecutter.project_name}}-datastore,DATA_STORE_REGION=us{%- elif cookiecutter.datastore_type == "vertex_ai_vector_search" %},VECTOR_SEARCH_INDEX={{cookiecutter.project_name}}-vector-search,VECTOR_SEARCH_INDEX_ENDPOINT={{cookiecutter.project_name}}-vector-search-endpoint,VECTOR_SEARCH_BUCKET=$$PROJECT_ID-{{cookiecutter.project_name}}-vs{%- endif %}{%- endif %}" \
+		$(if $(IAP),--iap) \
+		$(if $(PORT),--port=$(PORT))
 {%- elif cookiecutter.deployment_target == 'agent_engine' %}
 	# Export dependencies to requirements file using uv export.
-	uv export --no-hashes --no-header --no-dev --no-emit-project --no-annotate --frozen > .requirements.txt 2>/dev/null || \
-	uv export --no-hashes --no-header --no-dev --no-emit-project --frozen > .requirements.txt && uv run app/agent_engine_app.py
+	uv export --no-hashes --no-header --no-dev --no-emit-project --no-annotate > .requirements.txt 2>/dev/null || \
+	uv export --no-hashes --no-header --no-dev --no-emit-project > .requirements.txt && uv run app/agent_engine_app.py
 {%- endif %}
 {%- if cookiecutter.deployment_target == 'cloud_run' %}
 
+# Launch local development server with hot-reload
 local-backend:
 	uv run uvicorn app.server:app --host 0.0.0.0 --port 8000 --reload
 {%- endif %}
 {%- if cookiecutter.deployment_target == 'cloud_run' %}
 {%- if cookiecutter.agent_name == 'live_api' %}
 
+# Start the frontend UI for development
 ui:
 	PORT=8501 npm --prefix frontend start
 {%- endif %}
 {%- endif %}
 
+# Set up development environment resources using Terraform
 setup-dev-env:
 	PROJECT_ID=$$(gcloud config get-value project) && \
 	(cd deployment/terraform/dev && terraform init && terraform apply --var-file vars/env.tfvars --var dev_project_id=$$PROJECT_ID --auto-approve)
 
 {%- if cookiecutter.data_ingestion %}
 
+# Run the data ingestion pipeline for RAG capabilities
 data-ingestion:
 	PROJECT_ID=$$(gcloud config get-value project) && \
 	(cd data_ingestion && uv run data_ingestion_pipeline/submit_pipeline.py \
@@ -98,9 +111,11 @@ data-ingestion:
 		--pipeline-name="data-ingestion-pipeline")
 {%- endif %}
 
+# Run unit and integration tests
 test:
 	uv run pytest tests/unit && uv run pytest tests/integration
 
+# Run code quality checks (codespell, ruff, mypy)
 lint:
 	uv run codespell
 	uv run ruff check . --diff

src/base_template/deployment/cd/deploy-to-prod.yaml

@@ -21,7 +21,7 @@ steps:
       - -c
       - |
         cd data_ingestion && pip install uv==0.6.12 --user && cd data_ingestion_pipeline && \
-        uv sync --frozen && uv run python submit_pipeline.py
+        uv sync --locked && uv run python submit_pipeline.py
     env:
       - "PIPELINE_ROOT=${_PIPELINE_GCS_ROOT}"
       - "REGION=${_REGION}"
@@ -78,7 +78,7 @@ steps:
     args:
       - "-c"
       - |
-        pip install uv==0.6.12 --user && uv sync --frozen
+        pip install uv==0.6.12 --user && uv sync --locked
     env:
       - 'PATH=/usr/local/bin:/usr/bin:~/.local/bin'
 
@@ -88,7 +88,7 @@ steps:
     args:
       - "-c"
       - |
-        uv export --no-hashes --no-sources --no-header --no-dev --no-emit-project --no-annotate --frozen > .requirements.txt
+        uv export --no-hashes --no-sources --no-header --no-dev --no-emit-project --no-annotate --locked > .requirements.txt
         uv run app/agent_engine_app.py \
           --project ${_PROD_PROJECT_ID} \
           --location ${_REGION} \

src/base_template/deployment/cd/staging.yaml

@@ -21,7 +21,7 @@ steps:
       - -c
       - |
         cd data_ingestion && pip install uv==0.6.12 --user && cd data_ingestion_pipeline && \
-        uv sync --frozen && uv run python submit_pipeline.py
+        uv sync --locked && uv run python submit_pipeline.py
     env:
       - "PIPELINE_ROOT=${_PIPELINE_GCS_ROOT}"
       - "REGION=${_REGION}"
@@ -111,7 +111,7 @@ steps:
     args:
       - "-c"
       - |
-        pip install uv==0.6.12 --user && uv sync --frozen
+        pip install uv==0.6.12 --user && uv sync --locked
     env:
       - 'PATH=/usr/local/bin:/usr/bin:~/.local/bin'
 
@@ -121,7 +121,7 @@ steps:
     args:
       - "-c"
       - |
-        uv export --no-hashes --no-sources --no-header --no-dev --no-emit-project --no-annotate --frozen > .requirements.txt
+        uv export --no-hashes --no-sources --no-header --no-dev --no-emit-project --no-annotate --locked > .requirements.txt
         uv run app/agent_engine_app.py \
           --project ${_STAGING_PROJECT_ID} \
           --location ${_REGION} \
@@ -149,7 +149,7 @@ steps:
 {%- if cookiecutter.deployment_target == 'cloud_run' %}
         export _ID_TOKEN=$(cat id_token.txt)
         export _STAGING_URL=$(cat staging_url.txt)
-        pip install uv==0.6.12 --user && uv sync --frozen
+        pip install uv==0.6.12 --user && uv sync --locked
 {%- elif cookiecutter.deployment_target == 'agent_engine' %}
         export _AUTH_TOKEN=$(cat auth_token.txt)
 {%- endif %}

src/base_template/deployment/ci/pr_checks.yaml

@@ -20,7 +20,7 @@ steps:
     args:
       - "-c"
       - |
-        pip install uv==0.6.12 --user && uv sync --frozen
+        pip install uv==0.6.12 --user && uv sync --locked
     env:
       - 'PATH=/usr/local/bin:/usr/bin:~/.local/bin'
 

src/base_template/pyproject.toml

@@ -16,11 +16,11 @@ dependencies = [
 {%- endif %}
     "google-cloud-logging~=3.11.4",
 {%- if cookiecutter.deployment_target == 'cloud_run' %}
-    "google-cloud-aiplatform[evaluation]~=1.95.1",
+    "google-cloud-aiplatform[evaluation]~=1.99.0",
     "fastapi~=0.115.8",
     "uvicorn~=0.34.0"
 {%- elif cookiecutter.deployment_target == 'agent_engine' %}
-    "google-cloud-aiplatform[evaluation,agent-engines]~=1.95.1"
+    "google-cloud-aiplatform[evaluation,agent-engines]~=1.99.0"
 {%- endif %}
 ]
 {% if cookiecutter.deployment_target == 'cloud_run' %}

src/cli/utils/template.py

@@ -12,6 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import json
 import logging
 import os
 import pathlib
@@ -23,7 +24,7 @@ from typing import Any
 import yaml
 from cookiecutter.main import cookiecutter
 from rich.console import Console
-from rich.prompt import Prompt
+from rich.prompt import IntPrompt, Prompt
 
 from src.cli.utils.version import get_current_version
 
@@ -197,8 +198,6 @@ def prompt_deployment_target(agent_name: str) -> str:
         description = info.get("description", "")
         console.print(f"{idx}. [bold]{display_name}[/] - [dim]{description}[/]")
 
-    from rich.prompt import IntPrompt
-
     choice = IntPrompt.ask(
         "\nEnter the number of your deployment target choice",
         default=1,
@@ -510,6 +509,22 @@ def process_template(
             frontend_type = settings.get("frontend_type", DEFAULT_FRONTEND)
             tags = settings.get("tags", ["None"])
 
+            # Load adk-cheatsheet.md and llm.txt for injection
+            adk_cheatsheet_path = (
+                pathlib.Path(__file__).parent.parent.parent
+                / "resources"
+                / "docs"
+                / "adk-cheatsheet.md"
+            )
+            with open(adk_cheatsheet_path) as f:
+                adk_cheatsheet_content = f.read()
+
+            llm_txt_path = (
+                pathlib.Path(__file__).parent.parent.parent.parent / "llm.txt"
+            )
+            with open(llm_txt_path) as f:
+                llm_txt_content = f.read()
+
             cookiecutter_config = {
                 "project_name": "my-project",
                 "agent_name": agent_name,
@@ -525,11 +540,12 @@ def process_template(
                 "extra_dependencies": [extra_deps],
                 "data_ingestion": include_data_ingestion,
                 "datastore_type": datastore if datastore else "",
+                "adk_cheatsheet": adk_cheatsheet_content,
+                "llm_txt": llm_txt_content,
                 "_copy_without_render": [
                     "*.ipynb",  # Don't render notebooks
                     "*.json",  # Don't render JSON files
                     "frontend/*",  # Don't render frontend directory
-                    # "tests/*",  # Don't render tests directory
                     "notebooks/*",  # Don't render notebooks directory
                     ".git/*",  # Don't render git directory
                     "__pycache__/*",  # Don't render cache
@@ -537,15 +553,12 @@ def process_template(
                     ".pytest_cache/*",
                     ".venv/*",
                     "*templates.py",  # Don't render templates files
-                    "!*.py",  # render Python files
-                    "!Makefile",  # DO render Makefile
-                    "!README.md",  # DO render README.md
+                    # Don't render agent.py unless it's agentic_rag
+                    "app/agent.py" if agent_name != "agentic_rag" else "",
                 ],
             }
 
             with open(cookiecutter_template / "cookiecutter.json", "w") as f:
-                import json
-
                 json.dump(cookiecutter_config, f, indent=4)
 
             logging.debug(f"Template structure created at {cookiecutter_template}")