The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the agent to fetch documentation from an external URL: https://raw.githubusercontent.com/rustfs/rustfs/main/README.md. While the content is expected to be static markdown, relying on third-party hosted content introduces a dependency on external integrity.
[INDIRECT_PROMPT_INJECTION]: The skill implements a pattern where the AI agent ingests untrusted data from a remote repository via the WebFetch tool. An attacker who gains control of the specified GitHub repository could modify the README.md to include malicious instructions designed to influence the agent's behavior during deployment or configuration tasks.
Ingestion points: The WebFetch tool call located in the main instruction block of SKILL.md used to retrieve live documentation.
Boundary markers: Absent. The instructions do not provide delimiters or warnings to the agent to ignore instructions embedded within the fetched document.
Capability inventory: The skill is designed to facilitate the generation of configuration files, Docker commands, and API interactions for S3 storage, which are sensitive operations.
Sanitization: Absent. The skill does not specify any validation or filtering requirements for the content retrieved from the external source.

rustfs