The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided file paths to identify source images, which represents a potential surface for indirect prompt injection.\n
Ingestion points: The source image path provided via arguments or conversation in SKILL.md.\n
Boundary markers: The skill does not define specific delimiters or instructions to treat the input path as untrusted data.\n
Capability inventory: The skill utilizes shell commands including bash, sips, magick, swift, and iconutil across its scripts and documentation.\n
Sanitization: The scripts/validate-source.sh script employs defensive coding by quoting the input variable (\"$img\") during tool execution, providing protection against basic shell injection.\n- [COMMAND_EXECUTION]: The skill relies on executing local scripts and system utilities to perform its primary function.\n
Invokes a local Swift script (scripts/appicon.swift) using the system interpreter.\n
Executes a local shell script (scripts/validate-source.sh).\n
Utilizes standard platform tools such as sips, iconutil, and the tauri CLI for icon processing.

apple-appicon