circleci

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md and Reference workflow explicitly tell the agent to refresh and crawl public CircleCI documentation (e.g., "npx ctx7 docs /circleci/circleci-docs" and "$cf-crawl" / references/cf-crawl artifacts), so the agent will fetch and ingest open/public third‑party web pages and use them to drive recommendations and actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly requires running the Context7 CLI at runtime (e.g., npx ctx7 ...) and points to https://context7.com/docs/clients/cli — npx ctx7 will fetch and execute remote package code and retrieve content that the skill relies on to generate/update prompts/instructions, so this external URL/CLI dependency directly controls runtime behavior.