Rubenszimbres A2A Adk Mcp — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Rubenszimbres A2A Adk Mcp (Agent Skill) and scored it 82/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 2 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 2 flagged
The text {match} is the classic direct prompt-injection phrasing. Placed in a skill body that the agent reads as trusted instructions, it tries to make the agent abandon its prior rules and follow whatever comes next — a full system-prompt override.
ignore/disregard/forget … previous instructions sentence.The text {match} is the classic direct prompt-injection phrasing. Placed in a skill body that the agent reads as trusted instructions, it tries to make the agent abandon its prior rules and follow whatever comes next — a full system-prompt override.
ignore/disregard/forget … previous instructions sentence.Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
This project implements a security-focused data processing pipeline using the Agent-to-Agent (A2A) protocol combined with Model Context Protocol (MCP) integration. The system provides secure database queries through a multi-agent architecture.
The system utilizes a layered architecture:
The system implements a security pipeline with three specialized agents:
Client Request → A2A Server → Judge Agent → SQL Agent → Mask Agent → Client Responsepip install -r requirements.txtRun the adk web script to run the chat interface:
adk webThis will start:
Run the query_MCP_ADK_A2A.py script to query the multi-agent system:
python ./clients/query_MCP_ADK_A2A.pyThis will use the a2a_client.py module to make requests to the pipeline:
a2a_client.py: Client for A2A communicationa2a_servers.py: Server implementations for A2A protocolquery_MCP_ADK_A2A.py: Main pipeline implementationrun_servers.py: Server startup and coordinationserver_mcp.py: MCP server implementationtask_manager.py: Task coordination for agent communicationmcp_agent.py: Integration between ADK and MCPThe system integrates with Model Context Protocol (MCP) for enhanced SQL capabilities:
# Connect to MCP server
tools, exit_stack = await MCPToolset.from_server(
connection_params=StdioServerParameters(
command='python',
args=["server_mcp.py"],
)
)
# Create ADK agent with MCP tools
agent = LlmAgent(
model='gemini-2.5-pro-preview-03-25',
name='sql_assistant',
instruction="...",
tools=tools,
)Testing:
docker build -t adk-multi-agent .
docker run -p 8000:8000 -e GOOGLE_API_KEY=your_api_key adk-multi-agent adk webProduction:
export GOOGLE_CLOUD_PROJECT=next-project25
export GOOGLE_CLOUD_LOCATION=us-central1
export GOOGLE_GENAI_USE_VERTEXAI=True
export AGENT_PATH="."
export SERVICE_NAME="adk-agent-service"
export APP_NAME="agents"adk deploy cloud_run \
--project=$GOOGLE_CLOUD_PROJECT \
--region=$GOOGLE_CLOUD_LOCATION \
--service_name=$SERVICE_NAME \
--app_name=$APP_NAME \
--with_ui \
$AGENT_PATHAgent Development Kit Documentation
This project demonstrates integration between A2A protocol and MCP server capabilities, creating a secure and flexible agent architecture for data processing.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.