The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests user content in Step 1 to generate the infographic brief. While it does not utilize explicit boundary markers or delimiters to protect against instructions embedded within the user text, the skill's lack of high-risk capabilities (such as shell access or network connectivity) minimizes the potential impact of an indirect prompt injection attack.- [SAFE]: A thorough review of the skill instructions and metadata found no evidence of malicious patterns, obfuscation, or unauthorized data access. The use of optional local files (brand-kit.md, colours.md) for brand alignment is standard practice for this type of utility.

gemini-infographic