The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources such as XML sitemaps and text-based URL lists, which constitutes an indirect prompt injection surface.
Ingestion points: The scripts scripts/sitemap_analyzer.py, scripts/redirect_checker.py, and scripts/link_mapper.py read user-provided files (e.g., sitemap.xml, urls.txt).
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the processing logic.
Capability inventory: The skill uses subprocesses to run its own analysis scripts and scripts/redirect_checker.py performs network HEAD requests.
Sanitization: Content is processed using standard XML and URL parsers, though no specific filtering for natural language instructions within the data is implemented.
[COMMAND_EXECUTION]: The skill instructions include the execution of local Python scripts to perform site audits. These scripts are part of the skill package, use standard Python libraries, and are focused on processing provided input files to generate JSON or text reports.
[DATA_EXFILTRATION]: The script scripts/redirect_checker.py initiates network connections using http.client. It sends HEAD requests to URLs provided in the input file to verify HTTP status codes and redirect locations. This is the intended primary purpose of the redirect checker and does not involve the exfiltration of sensitive local files or credentials.

site-architecture