The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands such as xcrun, adb, kill, ffprobe, and ffmpeg to manage screen recording and verify media artifacts during mobile application testing. It also references running project-specific Dart tools via dart run.
[INDIRECT_PROMPT_INJECTION]: The skill processes external inputs such as 'requested behavior' and 'user-reported defects' to drive its testing logic, which could be used as vectors for indirect prompt injection. Ingestion points: Journey definitions and defect reports described in SKILL.md. Boundary markers: The instructions do not define specific markers or delimiters to isolate user-supplied test data from agent instructions. Capability inventory: The agent is granted access to system CLI tools and Dart execution as detailed in references/flutter.md. Sanitization: No specific methods for sanitizing or escaping the user-provided journey data are described in the skill.

e2e