The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run local Python scripts using shell commands to process uploaded files.
Evidence: python scripts/extract_structure.py /mnt/user-data/uploads/<filename>.xlsx and python scripts/extract_formulas.py /mnt/user-data/uploads/<filename>.xlsx found in SKILL.md.
Risk: The direct interpolation of a filename into a shell command template could lead to command injection if the file name is crafted with shell metacharacters (e.g., ;, |, &).
[PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data (Excel files), which introduces a surface for indirect prompt injection attacks.
Ingestion points: External data enters the agent context via the extraction scripts processing files at /mnt/user-data/uploads/<filename>.xlsx.
Boundary markers: Absent. There are no instructions or delimiters provided to ensure the agent ignores or isolates instructions that might be embedded within the Excel cell content, sheet names, or metadata.
Capability inventory: The skill possesses the ability to execute shell commands (python scripts/...) and generate local files.
Sanitization: No sanitization or validation logic is defined in the instructions to filter potential injection payloads from the spreadsheet data.

excel-auditor