The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute various shell commands for project auditing and performance analysis. This includes rg for code searching, xcodebuild for building projects, and system utilities such as ps, pgrep, lsof, sample, and vmmap for process monitoring and memory analysis.
[DATA_EXFILTRATION]: The skill provides instructions for the agent to access local application data paths, specifically $HOME/Library/Application Support/<AppName>/db.sqlite. While intended for diagnostic purposes (e.g., counting records for performance analysis), this requires accessing potentially sensitive user information stored in local databases.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to process user-provided source code and performance reports while having access to command execution and file system tools.
Ingestion points: User-provided Swift source code and performance reports (specified in references/ios-swiftui-entry-list-performance.md).
Boundary markers: The instructions do not define boundary markers or "ignore instructions" warnings for the external data being processed.
Capability inventory: The agent is granted capabilities to execute shell commands (sqlite3, xcodebuild, etc.) and read local system and application files.
Sanitization: There are no instructions provided for sanitizing or validating user-provided data before it is incorporated into diagnostic workflows.

yistc-swift