feat: Add Agent Skills standard support for ADK

contributing/samples/agent_skills_demo/README.md

@@ -0,0 +1,127 @@
+# Agent Skills Demo
+
+This demo showcases the Agent Skills standard integration with ADK. It demonstrates how to:
+
+1. Load skills from directories using `AgentSkillLoader`
+2. Convert skills to ADK tools using `SkillTool`
+3. Generate discovery prompts for the LLM
+4. Use skills with progressive disclosure
+
+## Running the Demo
+
+```bash
+# From the adk-python root directory
+adk web contributing/samples/agent_skills_demo
+```
+
+## What's Included
+
+### Skills Loaded
+
+The demo loads BigQuery ML skills from `src/google/adk/tools/bigquery/skills/`:
+
+- **bqml** - Machine learning in BigQuery
+- **bq-ai-operator** - AI operations (text generation, embeddings)
+
+### Skill Structure (Agent Skills Standard)
+
+Each skill follows the Agent Skills standard format:
+
+```
+skill-name/
+├── SKILL.md           # Short description and instructions
+├── references/        # Detailed documentation
+│   ├── MODEL_TYPES.md
+│   └── BEST_PRACTICES.md
+├── scripts/          # Helper scripts
+│   └── validate_model.py
+└── assets/           # Templates, configs
+```
+
+### Progressive Disclosure
+
+Skills support three stages:
+
+1. **Discovery** (Stage 1): Minimal metadata shown in discovery prompt
+2. **Activation** (Stage 2): Full instructions loaded on demand
+3. **Execution** (Stage 3): Scripts and references loaded as needed
+
+## Example Interactions
+
+### Discover Available Skills
+
+```
+User: What ML skills do you have?
+Agent: [Reviews discovery prompt and lists available skills]
+```
+
+### Activate a Skill
+
+```
+User: Tell me about BQML
+Agent: [Activates bqml skill, provides detailed instructions]
+```
+
+### Load Reference Documentation
+
+```
+User: What model types are available?
+Agent: [Loads MODEL_TYPES.md reference, explains options]
+```
+
+### Run a Script
+
+```
+User: Validate my model configuration
+Agent: [Runs validate_model.py script]
+```
+
+## Code Walkthrough
+
+```python
+from google.adk.skills import AgentSkillLoader, SkillTool
+
+# Load skills from directory
+loader = AgentSkillLoader()
+loader.add_skill_directory("./skills")
+
+# Create tools for agent
+skill_tools = [SkillTool(skill) for skill in loader.get_all_skills()]
+
+# Generate system prompt
+discovery_prompt = loader.generate_discovery_prompt()
+
+# Create agent with skill tools
+agent = LlmAgent(
+    model="gemini-2.0-flash",
+    instruction=f"Available skills:\n{discovery_prompt}",
+    tools=skill_tools,
+)
+```
+
+## Customization
+
+### Adding Custom Skills
+
+1. Create a skill directory following the Agent Skills standard
+2. Add `SKILL.md` with YAML frontmatter
+3. Add references, scripts, and assets as needed
+4. Update the `SKILLS_DIR` path in `agent.py`
+
+### Skill Configuration
+
+Skills can include ADK-specific configuration in their frontmatter:
+
+```yaml
+adk:
+  config:
+    timeout_seconds: 300
+    max_parallel_calls: 5
+  allowed_callers:
+    - my_agent
+```
+
+## Learn More
+
+- [Agent Skills Standard](https://agentskills.io)
+- [ADK Skills Documentation](../../docs/skills_programmatic_tool_calling_design.md)

contributing/samples/agent_skills_demo/__init__.py

@@ -0,0 +1,17 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Agent Skills Demo - demonstrates Agent Skills standard integration."""
+
+from . import agent

contributing/samples/agent_skills_demo/agent.py

@@ -0,0 +1,163 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Agent Skills Demo - demonstrates Agent Skills standard integration.
+
+This demo shows how to:
+1. Load skills from directories using AgentSkillLoader
+2. Convert skills to ADK tools using SkillTool
+3. Generate discovery prompts for the LLM
+4. Use skills with progressive disclosure
+5. Integrate with BigQuery toolset for data operations
+
+Run with: adk web contributing/samples
+Then select 'agent_skills_demo' from the list.
+"""
+
+from __future__ import annotations
+
+import os
+from pathlib import Path
+
+import google.auth
+
+from google.adk.agents.llm_agent import LlmAgent
+from google.adk.skills import AgentSkillLoader
+from google.adk.skills import SkillTool
+from google.adk.tools.bigquery.bigquery_credentials import BigQueryCredentialsConfig
+from google.adk.tools.bigquery.bigquery_toolset import BigQueryToolset
+from google.adk.tools.bigquery.config import BigQueryToolConfig
+from google.adk.tools.bigquery.config import WriteMode
+
+# Default Vertex AI settings
+os.environ.setdefault("GOOGLE_GENAI_USE_VERTEXAI", "true")
+os.environ.setdefault("GOOGLE_CLOUD_PROJECT", os.getenv("VERTEXAI_PROJECT", ""))
+os.environ.setdefault("GOOGLE_CLOUD_LOCATION", os.getenv("VERTEXAI_LOCATION", "us-central1"))
+
+# Path to skills directory (now in main skills module)
+SKILLS_DIR = Path(__file__).parent.parent.parent.parent / "src" / "google" / "adk" / "skills"
+
+# Initialize the skill loader
+skill_loader = AgentSkillLoader(validate_names=False)
+
+# Load skills from the BigQuery skills directory
+if SKILLS_DIR.exists():
+    skill_count = skill_loader.add_skill_directory(SKILLS_DIR)
+    print(f"Loaded {skill_count} skills from {SKILLS_DIR}")
+else:
+    print(f"Warning: Skills directory not found: {SKILLS_DIR}")
+
+# Create skill tools for the agent
+skill_tools = [SkillTool(skill) for skill in skill_loader.get_all_skills()]
+
+# Generate discovery prompt for the system instruction
+discovery_prompt = skill_loader.generate_discovery_prompt(include_resources=True)
+
+# Setup BigQuery toolset with Application Default Credentials
+try:
+    application_default_credentials, _ = google.auth.default()
+    credentials_config = BigQueryCredentialsConfig(
+        credentials=application_default_credentials
+    )
+
+    # Configure BigQuery tools - allow writes for ML model creation
+    tool_config = BigQueryToolConfig(
+        write_mode=WriteMode.ALLOWED,
+        application_name="agent_skills_demo"
+    )
+
+    bigquery_toolset = BigQueryToolset(
+        credentials_config=credentials_config,
+        bigquery_tool_config=tool_config
+    )
+    print("BigQuery toolset initialized successfully")
+except Exception as e:
+    print(f"Warning: Could not initialize BigQuery toolset: {e}")
+    bigquery_toolset = None
+
+# Build the system instruction
+SYSTEM_INSTRUCTION = f"""You are a data science assistant with access to BigQuery AI and ML skills.
+
+## Available Skills
+
+{discovery_prompt}
+
+## Skill Overview
+
+- **bigquery-ai**: Generative AI operations - text generation with LLMs (Gemini, Claude), embeddings, vector search, and RAG workflows
+- **bqml**: Traditional ML - classification, regression, clustering, time series forecasting, recommendations
+
+## How to Use Skills
+
+1. **Discover**: Review the available skills above to find the right one for your task
+2. **Activate**: Use the skill tool with action="activate" to get detailed instructions
+3. **Load References**: Use action="load_reference" to load detailed documentation for specific topics
+4. **Run Scripts**: Use action="run_script" to execute helper scripts for setup and validation
+
+## BigQuery Tools
+
+You also have direct access to BigQuery tools for:
+- Executing SQL queries (including CREATE MODEL, ML functions)
+- Exploring datasets and tables
+- Getting table schemas and metadata
+
+## Guidelines
+
+- For generative AI (text generation, embeddings, semantic search, RAG): Use **bigquery-ai** skill
+- For predictive ML (classification, regression, forecasting): Use **bqml** skill
+- Always activate a skill before using its detailed features
+- Load specific reference docs when you need in-depth information
+- Use BigQuery tools to run the actual SQL queries
+
+## Example Workflows
+
+**Generative AI Example:**
+User: "How do I build a RAG system in BigQuery?"
+1. Activate bigquery-ai skill
+2. Load RAG_WORKFLOW.md reference
+3. Use BigQuery tools to create models and run queries
+
+**Traditional ML Example:**
+User: "How do I train a churn prediction model?"
+1. Activate bqml skill
+2. Load MODEL_TYPES.md reference for classifier options
+3. Use BigQuery tools to create and evaluate the model
+"""
+
+# Combine all tools
+all_tools = skill_tools.copy()
+if bigquery_toolset:
+    all_tools.append(bigquery_toolset)
+
+# Create the agent with Gemini 2.5 Pro
+root_agent = LlmAgent(
+    model="gemini-2.5-pro",
+    name="agent_skills_demo",
+    description="A demo agent showcasing Agent Skills standard integration with ADK and BigQuery tools.",
+    instruction=SYSTEM_INSTRUCTION,
+    tools=all_tools,
+)
+
+# Print summary on load
+if __name__ == "__main__" or True:
+    print("\n" + "=" * 60)
+    print("Agent Skills Demo")
+    print("=" * 60)
+    print(f"\nModel: gemini-2.5-pro")
+    print(f"Loaded skills: {skill_loader.get_skill_names()}")
+    print(f"Skill tools: {[t.name for t in skill_tools]}")
+    print(f"BigQuery toolset: {'enabled' if bigquery_toolset else 'disabled'}")
+    print("\nRun with: adk web contributing/samples")
+    print("Then select 'agent_skills_demo' from the list")
+    print("=" * 60 + "\n")