test-mcp-server
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user to clone reference code from the author's official GitHub repository (Alignbase/sunpeak) to provide working test examples. This is documented neutrally as it targets the vendor's own infrastructure for legitimate resource sharing.
- [COMMAND_EXECUTION]: The skill utilizes a set of CLI commands (
sunpeak test,sunpeak inspect,sunpeak test init) to scaffold project infrastructure and execute test suites. These commands are standard for a development tool and are restricted to the testing and initialization tasks described in the documentation. - [INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection because it is designed to ingest and process data from external MCP servers and live LLM interactions during testing.
- Ingestion points: Data enters the context via
inspector.renderTool,mcp.callTool, andlive.invokewhen interacting with external MCP tools or live chat sessions. - Boundary markers: None explicitly defined in the provided snippets, though standard for testing tools of this nature.
- Capability inventory: The skill has the capability to control browsers (via Playwright) and execute local shell commands for test execution.
- Sanitization: Not explicitly documented; however, as a testing framework, it is intended to evaluate the behavior of the data it ingests rather than execute it blindly.
Audit Metadata