The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands including find, ls, du, mkdir, and mv to analyze and modify the file system. While these are required for its stated purpose, they carry inherent risks if target directories contain sensitive system files or if inputs are not properly sanitized by the agent.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8). 1. Ingestion points: File names and content are read via ls -la, find -exec file, and find -printf in SKILL.md during the analysis phase. 2. Boundary markers: None identified; the agent processes file data directly into its context. 3. Capability inventory: The agent has the capability to create directories (mkdir) and move/rename files (mv) as described in SKILL.md. 4. Sanitization: No explicit sanitization or filtering of file names or content is defined before processing.

file-organizer