The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill identifies a vulnerability surface where untrusted data from the filesystem is ingested into the agent context.
Ingestion points: The skill reads directory content and filenames using ls, find, and du in SKILL.md. These paths (like the Downloads folder) frequently contain files from untrusted external sources.
Boundary markers: There are no explicit instructions or delimiters provided to ensure the agent ignores or sanitizes malicious instructions that might be embedded in filenames or file metadata.
Capability inventory: The skill possesses filesystem modification capabilities, specifically through mkdir and mv commands defined in SKILL.md.
Sanitization: The instructions lack explicit requirements for sanitizing or validating filenames before they are interpolated into shell commands, creating a potential vector for command injection via maliciously crafted filenames.
[COMMAND_EXECUTION]: The skill relies on local shell command execution to perform its primary functions.
Evidence: SKILL.md contains multiple bash snippets using ls, find, du, md5, mkdir, and mv to analyze and restructure the filesystem.

file-organizer