The Agent Skills Directory

[SAFE]: The skill implements structural validation for local project files. All operations, including file reading and frontmatter updates, are restricted to the local marketplace directory.
[COMMAND_EXECUTION]: The script executes a local git command (git remote get-url origin) to verify if the marketplace name matches the repository name. This invocation uses a fixed argument list, which is a safe implementation that prevents shell injection.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from local project files, such as marketplace.json and markdown frontmatter. While the script lacks explicit boundary markers in its output reports, its logic is limited to structural validation and regex parsing. There is no evidence of executable interpretation of this content, and the ingestion is central to the skill's primary auditing purpose.

marketplace-audit