The Agent Skills Directory

[SAFE]: No security concerns identified. The script uses only the Python standard library and operates strictly on files provided by the user.
[PROMPT_INJECTION]: The skill handles untrusted CSV data which presents an indirect prompt injection surface if the file contains instructions within model names or provider fields. However, the script primarily processes numeric data and lacks dangerous capabilities. * Ingestion points: scripts/analyze.py (line 33) * Boundary markers: Absent; script output is interpreted directly by the agent * Capability inventory: Local file reading and printing to standard output * Sanitization: Numeric fields are explicitly cast to float or integer types.

openrouter-analyzer