ccpi-marketplace

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

• Third-party content exposure detected (high risk: 0.70). The required workflow centers on running the ccpi CLI and Claude plugin marketplace commands (e.g., ccpi search <keyword> and plugin installs), which at runtime necessarily fetch/displays marketplace-provided pack metadata and/or search results that are outsider-authored free text; this text can be fed into the agent’s LLM context as part of outputs to reason over.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs installing and using the ccpi CLI and adding the jeremylongshore marketplace which will fetch and execute remote package/plugin code at runtime from https://www.npmjs.com/package/@intentsolutionsio/ccpi and https://github.com/jeremylongshore/claude-code-plugins-plus-skills, so these external sources can execute code or supply plugins that control agent behavior.