The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill directs users to run npx shadcn@latest add https://raw.githubusercontent.com/Aslam97/shadcn-minimal-tiptap/main/registry/block-registry.json. This fetches a component registry from a third-party GitHub repository that is not associated with a trusted organization, introducing unverifiable code into the project environment.\n- [COMMAND_EXECUTION]: The provided templates/minimal-tiptap-setup.sh script automates the installation of the third-party shadcn component and various npm packages. Executing shell scripts that incorporate remote resources presents a security risk if the remote source is compromised.\n- [INDIRECT_PROMPT_INJECTION]: The editor components are designed to process rich text content, which serves as an ingestion point for untrusted data.\n
Ingestion points: The content prop in templates/base-editor.tsx and the handlePaste/handleDrop logic in templates/image-upload-r2.tsx.\n
Boundary markers: There are no explicit instructions or code patterns provided to delimit untrusted content or warn the agent against executing instructions contained within the edited text.\n
Capability inventory: The skill includes templates for network requests (image uploads) and complex state management, which could be targeted by malicious inputs.\n
Sanitization: While the Tiptap framework itself provides some protection through its schema, the skill does not explicitly guide the user on sanitizing output to prevent downstream instruction injection.

tiptap