The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script scripts/analyzer.py to perform its analytical tasks. The script uses standard libraries and performs data processing without any network access, file system modifications, or secondary subprocess spawning.
[PROMPT_INJECTION]: The skill processes untrusted data in the form of product reviews (competitor reviews and user store reviews). This content is extracted and presented to the AI agent in the final analysis report, creating a potential surface for indirect prompt injection.
Ingestion points: Review text is ingested via the ReviewData class in scripts/analyzer.py and processed through keyword matching.
Boundary markers: The report generated by format_report wraps review examples in standard quotation marks but lacks explicit delimiters or instructions to the agent to disregard instructions embedded within the reviews.
Capability inventory: The skill code contains no capabilities for network operations, arbitrary command execution, or sensitive file access.
Sanitization: No content sanitization or instruction filtering is performed on the review text; it is only truncated to 100 characters before inclusion in the report.

product-differentiation-shopify