The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill includes a command /baseline-ui <file> that enables the ingestion of untrusted external file content into the agent context, creating a potential vector for malicious instructions to be processed by the model.
Ingestion points: File content provided via the <file> argument in the /baseline-ui command (SKILL.md).
Boundary markers: The instructions do not specify any delimiters or warnings to ignore instructions that might be embedded within the provided file content.
Capability inventory: The skill is capable of reading file content, identifying violations of its UI baseline, and generating code-level suggestions and fixes across the project.
Sanitization: No explicit sanitization, validation, or escaping of the external file content is described in the skill instructions.
[EXTERNAL_DOWNLOADS]: The skill documentation refers to shadcn-vue.com as a preferred source for component primitives. Automated scanners have flagged this specific domain and its documentation sub-paths as potentially malicious (identified as a CryptScam/URL Block). While shadcn-vue is a well-known open-source library in the Vue ecosystem, the reputation alert for this specific domain suggests users should verify the resource integrity before following the recommendation.

baseline-ui