The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is centered around the tauri-mcp CLI tool, which executes various subcommands in the shell environment. This includes starting driver sessions, managing daemons, and interacting with mobile devices via adb.
[DYNAMIC_EXECUTION]: The skill provides tools for executing arbitrary JavaScript within the application's webview using tauri-mcp webview-execute-js. It also allows executing backend commands through tauri-mcp ipc-execute-command. These capabilities allow for dynamic code execution within the context of the application being debugged.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection because it instructs the agent to process data from untrusted external sources.
Ingestion points: The agent reads potentially attacker-controlled data via tauri-mcp read-logs (which fetches console and system logs), tauri-mcp webview-execute-js (which returns DOM content), and tauri-mcp webview-find-element (which inspects UI state).
Boundary markers: Absent. The skill provides no instructions or delimiters to help the agent distinguish between its own system instructions and instructions that may be embedded in the logs or webview content it processes.
Capability inventory: The agent has the ability to execute shell commands (tauri-mcp), write files to disk (webview-screenshot --file), and perform UI interactions that could be leveraged if the agent is manipulated by injected data.
Sanitization: Absent. There are no mentions of escaping, filtering, or validating the content retrieved from the webview or logs before it is processed by the agent.

tauri-mcp-cli