The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes user-supplied brand briefs, documents, and URLs to generate campaign content. This creates a surface for indirect prompt injection where malicious instructions could be embedded in the source materials.
Ingestion points: Intake phase in SKILL.md (URLs, docs, text blocks).
Boundary markers: The prompts do not include explicit delimiters or instructions to ignore embedded commands within the ingested data.
Capability inventory: The agent creates directories and writes multiple markdown files to the local file system. It also utilizes ToolSearch to discover and invoke Higgsfield-related visual generation tools.
Sanitization: No explicit validation or filtering of user-provided text is performed before interpolation into the generation prompts.
[EXTERNAL_DOWNLOADS]: The skill references the Higgsfield MCP documentation and tools for rendering campaign visuals. This is a reference to a well-known AI visual generation service.

cmo-agent