The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions require the agent to download and execute the native-devtools-mcp package from the npm registry using npx or pnpm dlx. This is a standard procedure for the functionality of this skill.
[COMMAND_EXECUTION]: The skill directs the agent to execute shell commands to launch Electron applications with specific debugging flags and to verify network endpoints using curl on the local loopback interface (127.0.0.1).
[PROMPT_INJECTION]: The skill has a vulnerability surface for Indirect Prompt Injection (Category 8) because it retrieves and processes data from external application UIs that could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The agent reads DOM snapshots, OCR results from screenshots, and accessibility trees from the application under test (as described in SKILL.md).
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat UI content as untrusted or to ignore instructions embedded within it.
Capability inventory: The agent has the ability to execute shell commands (npm, electron, curl) and control the desktop environment via the MCP tool.
Sanitization: There is no evidence of sanitization or filtering of the content extracted from the application's UI before it is processed by the agent.

electron-live-test