The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes multiple internal documents (Core Documents 3.1-3.4, Part 4/7 documents, and the Living Project Instruction File) and synthesizes their content into a flagship deliverable.
Ingestion points: Reads files from ~/.claude-marketing/brands/{slug}/ and various path patterns like ../context-engine/.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when reading these files.
Capability inventory: The skill has access to Bash, Write, Edit, and Read tools.
Sanitization: There is no evidence of sanitization or escaping of the content read from these documents before synthesis.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute local scripts for document conversion.
Evidence: The production process invokes pdf-generator.py and other 'document export utilities' to generate PDF and DOCX files.
Risk: While these are intended internal utilities, executing scripts that process synthesized data could be leveraged if an indirect prompt injection attack successfully compromises the synthesis step.

growth-plan