context7-cli

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md and referenced docs (references/docs.md and references/skills.md) explicitly instruct the agent to run ctx7 library/ctx7 docs to fetch up-to-date library documentation and to install skills from public GitHub repos (e.g., ctx7 skills install /owner/repo), meaning the agent will ingest untrusted, user-generated third-party content that can influence its subsequent tool use and behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill allows runtime installation from GitHub-style repositories (e.g., ctx7 skills install /anthropics/skills), which fetches Markdown "skills" at runtime that are intended to teach or instruct AI agents and therefore can directly control agent prompts/instructions.