The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation and examples rely on npx mdbase, which triggers the download of the mdbase package from the public NPM registry during execution.
[REMOTE_CODE_EXECUTION]: By using npx to execute a package that is not pinned to a specific version and originates from a third-party developer (callumalpass) outside the trusted vendor list, the skill introduces a risk of remote code execution if the package is compromised.
[COMMAND_EXECUTION]: The skill involves the execution of various shell commands, including npx, ls, and xargs, to perform operations like querying, updating, and validating the local file system.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection (Category 8) because it reads and processes user-controlled markdown and YAML files.
Ingestion points: The npx mdbase query, npx mdbase read, and npx mdbase validate commands ingest data from local .md and .yaml files.
Boundary markers: The files use standard YAML frontmatter delimiters (---), but there are no specific instructions to the agent to disregard instructions found within the data fields or markdown body.
Capability inventory: The skill can execute shell commands, create/modify/rename files, and export data to external formats (CSV/JSON).
Sanitization: There is no evidence of input sanitization or validation to prevent the agent from acting on instructions embedded in the processed files.

tools-mdbase