The Agent Skills Directory

[DATA_EXPOSURE]: The skill incorporates a proactive security measure by instructing the agent to scan for and remove credentials, API keys, and secrets from any source content before it is processed or saved to output files.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests external content from users, it manages this risk through a structured analysis workflow that separates data from design instructions and emphasizes data integrity without executing embedded commands.
[SAFE]: The skill relies on local reference files for layout and style definitions and uses established agent tools (read_file, write_file, clarify, and image_generate) for its core operations within a defined workspace.

baoyu-infographic