quickstart-guide

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). At runtime the workflow extracts a Quickstart slug from user-provided URLs and then fetches the corresponding markdown content from GitHub (via the GitHub contents API and reading the raw .md download_url), so outsider-provided URLs can cause the agent to ingest free text authored in that repository.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill fetches and executes content at runtime from GitHub (e.g., https://api.github.com/repos/Snowflake-Labs/sfquickstarts/contents/site/sfguides/src/<slug> and companion repo URLs), which can supply markdown/code that the agent parses and executes (via sql_execute or git clone), so external content directly controls execution.