The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection by ingesting configuration and context from external files.
Ingestion points: Files include _active-brand.json, profile.json, and _manifest.json located in the ~/.claude-marketing/ directory, as well as compliance-rules.md.
Boundary markers: No delimiters or instructions are provided to distinguish between the skill's core instructions and the data loaded from external files.
Capability inventory: The skill is limited to analysis and reporting; it does not utilize tools for network communication, arbitrary code execution, or unauthorized file system modification.
Sanitization: There are no mechanisms described to validate or sanitize the input from the external context files before they are processed by the agent.

attribution-model