The Agent Skills Directory

[PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill processes untrusted data extracted from invoices and receipts which could contain malicious instructions.
Ingestion points: Reading content from PDF files and images (OCR/text extraction) as described in SKILL.md, Step 2.
Boundary markers: Absent; the instructions do not specify the use of delimiters or specific warnings to ignore instructions embedded within the extracted text.
Capability inventory: The skill uses shell commands including find, mkdir, cp, and mv to organize files based on extracted data.
Sanitization: Step 4 in SKILL.md instructs the agent to "Remove special characters except hyphens" from filenames, which provides a layer of protection against command injection via malicious file metadata.
[COMMAND_EXECUTION]: Uses local shell commands for file system management.
Evidence: The skill utilizes find to locate files and mkdir, cp, and mv to create the organizational structure and move files accordingly (SKILL.md, Steps 1 and 5). These commands are limited to local operations consistent with the skill's stated purpose.

invoice-organizer