The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches external configuration and infrastructure-as-code from a third-party GitHub repository (https://github.com/Jamonygr/Azure-Security-Lab.git). This repository is not identified as a trusted organization or the official vendor for this skill.
[REMOTE_CODE_EXECUTION]: Provides instructions to execute downloaded content using terraform apply. This pattern involves running unverified logic from an external source, which can lead to unauthorized infrastructure changes, resource creation, or credential exposure within the user's Azure subscription.
[COMMAND_EXECUTION]: Recommends the use of terraform apply -auto-approve, which bypasses the confirmation step and automatically deploys resources. This significantly reduces the opportunity for users to review the planned infrastructure changes for security risks or malicious configurations.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by processing and outputting data from CLI tools.
Ingestion points: External data enters the context through terraform output and az keyvault secret show commands.
Boundary markers: No delimiters or safety instructions are used to differentiate between trusted and untrusted data.
Capability inventory: The skill has broad capabilities including git, terraform, and az CLI access.
Sanitization: No validation or escaping is applied to the outputs retrieved from these tools before they are presented to the agent.

azure-security-lab-terraform