azure-hosted-copilot-sdk

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs fetching and reading public GitHub repositories (e.g., references/copilot-sdk.md and references/existing-project-integration.md direct the agent to call github-mcp-server-get_file_contents on public repos like "github/copilot-sdk" and "azure-samples/copilot-sdk-service"), which are untrusted third‑party sources whose content the agent is expected to interpret and use to drive deployment and code-change actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs runtime fetching of repository content that will be injected/used by the agent (e.g., "github-mcp-server-get_file_contents" to read https://github.com/github/copilot-sdk and azd init --template azure-samples/copilot-sdk-service which fetches the azure-samples template at https://github.com/azure-samples/copilot-sdk-service), so external repository content is fetched at runtime and can directly control prompts or supply executable scaffold/scripts.