The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The documentation in references/linux-wsl.md provides instructions to download an executable binary from the vendor's GitHub repository (github.com/xtool-org/xtool) using curl and making it executable.
[COMMAND_EXECUTION]: The skill uses a Python script (scripts/check-environment.py) that executes external binaries like xtool and swift via subprocess.run to verify the host environment.
[INDIRECT_PROMPT_INJECTION]: The skill processes and displays untrusted data from tool outputs and configuration files, creating a surface for indirect prompt injection.
Ingestion points: The agent reads output from xtool and swift commands via scripts/check-environment.py, and references project files like xtool.yml and Package.swift as described in SKILL.md and references/project-config.md.
Boundary markers: The instructions do not define clear delimiters or use 'ignore embedded instructions' warnings when processing this external content.
Capability inventory: The skill has the ability to execute shell commands and run subprocesses, which could be leveraged if an injection is successful.
Sanitization: The environment check script prints raw tool output without validation or escaping before it is seen by the agent context.

xtool