The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is centered around executing the 'marionette' CLI tool with various parameters to control Flutter applications. This includes registering instances, interacting with UI elements, and starting an MCP server.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes untrusted data from the application under test.
Ingestion points: The get-logs and get-interactive-elements commands in SKILL.md retrieve arbitrary text data from the target Flutter application.
Boundary markers: No boundary markers or instructions to ignore embedded commands are present in the skill definition.
Capability inventory: The agent has the ability to execute CLI commands (marionette) and write to the local filesystem using the take-screenshots --output command in SKILL.md.
Sanitization: There is no evidence of sanitization, filtering, or validation performed on the log entries or UI element text before they are processed by the agent.
[COMMAND_EXECUTION]: The take-screenshots command allows the agent to specify an arbitrary output file path via the --output flag. If manipulated, this could potentially be used to write files to sensitive locations accessible by the agent.

testing-with-marionette