braintrust

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

• Potentially malicious external URL detected (high risk: 0.80). The README shows runtime setup/fetch commands (e.g., bt docs fetch / bt setup prefetch) that sync Markdown from https://www.braintrust.dev/docs into LLM-friendly local indexes used by agent skills, so fetched content from that URL can directly influence agent prompts (see SKILL.md lines 190-201 and 223-225).