The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/install.sh and scripts/install.ps1 files contain commands to install the xml2 package using system package managers (apt, brew, dnf, etc.). While these scripts modify the system environment, they are standard installation procedures for the utility provided by the skill.
[EXTERNAL_DOWNLOADS]: Several examples in references/cheatsheet.md and references/examples-and-recipes.md demonstrate fetching XML data from external sources using curl. This is documented as a common use case for the tool and uses a well-known example domain (example.com).
[PROMPT_INJECTION]: As the skill is designed to ingest and process external XML data, it presents a surface for indirect prompt injection. A maliciously crafted XML file could attempt to influence the agent's interpretation when the flattened results are processed in a pipeline.
Ingestion points: XML data is read from files or stdin in SKILL.md and references/cheatsheet.md.
Boundary markers: None present in the instructions to delimit external data from agent commands.
Capability inventory: The skill utilizes xml2, 2xml, grep, sed, awk, xmllint, and xq for data processing and reconstruction.
Sanitization: No explicit sanitization or validation of the XML structure or content is performed before processing.

xml2