The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted XML data from construction systems, which represents a potential surface for indirect prompt injection or XML-based attacks.
Ingestion points: The parse_file and parse_string methods in SKILL.md accept external data inputs.
Boundary markers: No explicit delimiters or isolation instructions are implemented to separate processed data from agent instructions.
Capability inventory: The skill has filesystem read access to process project files.
Sanitization: The implementation uses standard xml.etree.ElementTree, which lacks specialized protection against certain complex XML payloads compared to libraries like defusedxml.
[SAFE]: All identified functionality and external references to datadrivenconstruction.io are consistent with the skill's stated purpose and represent legitimate vendor resources.

xml-reader