azure-hosted-copilot-sdk

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

• Third-party content exposure detected (high risk: 1.00). This skill explicitly instructs the agent to fetch and read public GitHub-hosted docs/code (e.g., "Use context7 MCP tools as the PRIMARY way to get SDK documentation and code examples" and the fallback "github-mcp-server-get_file_contents" in references/copilot-sdk.md), which are untrusted third-party, user-generated sources that the agent must interpret and that can influence 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 SDK documentation/snippets via context7 MCP calls and a fallback to reading the GitHub repo (e.g. https://github.com/github/copilot-sdk or via github-mcp-server-get_file_contents owner:"github", repo:"copilot-sdk"), which would inject externally-fetched content into the agent context and thus directly control prompts/instructions.