notion-skills-usage

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

• Third-party content exposure detected (high risk: 0.75). SKILL.md describes install/sync/feed workflows that pull skill content from Notion pages (outsider-authored by other people in the team store) into on-disk SKILL.md/references//scripts/, which the agent can then load into its LLM context—indirect prompt injection risk via Notion-fetched free text.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly supports pulling public GitHub repos at runtime (e.g., "notion-skills add vercel-labs/agent-skills"), which will fetch remote SKILL.md and any scripts from those repos that can contain instructions controlling agent prompts or executable code, so this is a runtime external dependency that meets the flagging criteria.