The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from various project files and external landing pages to generate marketing plans, creating a potential surface for indirect prompt injection attacks. * Ingestion points: The skill reads project-root files including CLAUDE.md, README.md, package.json, and 'landing pages' (Phase 0). * Boundary markers: The instructions lack explicit delimiters or 'ignore embedded instructions' warnings for the data being processed. * Capability inventory: The skill has the capability to write new files (MARKETING.md) to the project root directory. * Sanitization: There is no mention of sanitization, validation, or filtering of the content extracted from project files or external sources before it is used in the model or written to disk.

kai-budget