The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill identifies the 'xmind' Python package as a dependency and instructs the agent to install it via pip if it is not present. This is a standard library for processing XMind documents and is considered a safe external reference in this context.
[COMMAND_EXECUTION]: The skill utilizes PowerShell commands such as Test-Path and Expand-Archive, as well as Python's zipfile and os modules, to manipulate and read local files. These operations are scoped to the file paths provided by the user during interaction.
[PROMPT_INJECTION]: The skill processes content from external .xmind files, creating a vulnerability surface for indirect prompt injection where malicious instructions in a file could influence the AI. 1. Ingestion points: The read_xmind and xmind_to_markdown functions in SKILL.md ingest data from external files. 2. Boundary markers: The skill does not implement delimiters or warnings to ignore instructions embedded within the mind map content. 3. Capability inventory: The agent has the ability to read/write files and install packages. 4. Sanitization: No evidence of escaping or filtering content extracted from the XMind nodes before presentation to the AI.

xmind-reader