The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (LOW): The skill performs network requests to mlflow.org to fetch documentation. While this is the intended purpose, it is a non-whitelisted external domain.
[PROMPT_INJECTION] (LOW): The skill is vulnerable to indirect prompt injection (Category 8) because it ingests untrusted data from an external website and incorporates it into the agent's context.
Ingestion points: The skill fetches the llms.txt index and various .md documentation files from https://mlflow.org/docs/latest/.
Boundary markers: Absent. The instructions do not use delimiters or warnings to ignore instructions that might be embedded in the documentation.
Capability inventory: The skill uses WebFetch to read external data. While the skill itself does not define write or execute capabilities, the data is returned to the agent which may have such tools.
Sanitization: No sanitization, escaping, or validation of the fetched documentation content is performed before presentation.

searching-mlflow-docs