The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data to generate diagrams, which creates a surface for indirect instructions to influence agent behavior.
Ingestion points: The skill ingests user-provided text descriptions in chat, local source code referenced via the sources[] array, and file paths specified in data-src attributes.
Boundary markers: There are no explicit delimiters or system instructions defined to treat the ingested data as untrusted or to ignore embedded instructions.
Capability inventory: The skill can execute shell commands via Node.js and Python and has read access to the local file system within the configured repository root.
Sanitization: The skill employs automated validation tools (frontend-antipattern.py and visual-receipt.py) to check for geometric errors and accessibility, but these do not perform security-focused sanitization of the input content.
[EXTERNAL_DOWNLOADS]: The skill provides instructions to download and install a specific external component from a vendor repository using the command npx skills add Rheinmir/archify.
[COMMAND_EXECUTION]: The skill relies on several local shell commands to function, including running the diagram engine via node and performing post-processing and validation using python3 scripts.

diagram