The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted MTHDS bundle files, which introduces an indirect prompt injection surface where malicious instructions could be embedded in the data being explained or executed.
Ingestion points: Reads .mthds bundle files in Step 1.
Boundary markers: The instructions do not define delimiters or provide specific warnings to the model to ignore instructions found within the bundle data.
Capability inventory: The skill can execute CLI commands (mthds-agent) and generate local files (live_run.html), which could be leveraged if a bundle is successfully used to inject instructions.
Sanitization: There is no evidence of content sanitization or validation before the information is presented to the user or passed to the CLI tool.
[COMMAND_EXECUTION]: The skill invokes the mthds-agent CLI for version checks, bundle validation, and execution. These commands are part of the core functionality and utilize tools provided by the vendor 'mthds-ai'.
[EXTERNAL_DOWNLOADS]: The skill recommends installing the mthds package from the npm registry. This is the official package for the MTHDS platform and is considered a safe vendor resource.

mthds-explain