The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill acts as an installer that prompts the agent to download and set up a full toolset from a third-party GitHub repository (github.com/Nutlope/hallmark) using the npx skills add command.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface through its audit and study verbs, which are designed to process external URLs and screenshots. This creates a risk where malicious instructions embedded in a target's website content or image metadata could influence the agent's behavior during analysis.
Ingestion points: Untrusted URLs and screenshots provided to the hallmark audit and hallmark study commands as described in SKILL.md.
Boundary markers: No explicit delimiters or instructions to ignore embedded content are provided in the skill stub.
Capability inventory: The skill is capable of performing structural audits, visual redesigns, and design DNA extraction, which may involve reading external content and writing files (e.g., design.md).
Sanitization: No input validation or sanitization mechanisms are defined in the analyzed files.

hallmark