cargo-context

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

• Third-party content exposure detected (high risk: 0.90). The skill's bootstrap workflow (references/examples/bootstrap-and-update.md and SKILL.md) explicitly instructs sub-agents to scrape and ingest public sources — e.g., websites, review sites, Reddit/Hacker News — and use those digests to seed the context repo that agents read and act on, exposing the agent to untrusted, user-generated third-party content that can influence subsequent actions.