The TEE Plugin is a plugin designed to obtain an attestation report in a Trusted Execution Environment (TEE).
An attestation report is a document generated by a Trusted Execution Environment (TEE) that serves as cryptographic proof of the environment's integrity and trustworthiness. It is primarily used to assure external parties that:
-
The TEE environment is secure: It proves that the execution environment is isolated from other processes and protected from unauthorized access or tampering.
-
The code running inside the TEE is legitimate: It verifies that the code or application executed inside the TEE has not been altered and is authentic.
-
The environment has not been compromised: It includes evidence that the hardware and software configurations are intact and match a known, trusted state.
From this directory (tee), run the installation:
pip install -e .This TEE plugin currently supports retrieving attestation reports exclusively from Google Confidential Space. It is actively under development, and support for attestation reports from AMD SEV-SNP Confidential VMs will be released soon.
- You should deploy the AI Agent on Google Confidential Space to safeguard sensitive information, such as wallet secrets.
- Use this plugin to generate an attestation report for Google Confidential Space, demonstrating that the agent is enhanced by TEE and operating in a secure environment. The attestation report also includes the agent's Docker image to verify that it matches the expected configuration.
- Import and initialize the plugin to use in your worker:
from tee_plugin_gamesdk.tee_plugin import TeePlugin
options = {
"id": "test_tee_worker",
"name": "Test TEE Worker",
"description": "An example TEE Plugin for testing.",
"type": "GCS"
}
# Initialize the TeePlugin with your options
tee_plugin = TeePlugin(options)
# Generate Attestation report
get_attestation_report_fn = tee_plugin.get_function('get_attestation_report')
get_attestation_report_fn("Hello world!") # The input is a nonce releated to the reportYou can refer to test_tee.py for more examples on how to call the twitter functions.