The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its reliance on external data sources for context construction. Ingestion points: The agent reads brand profiles, compliance rules, manifest files, and custom templates from the ~/.claude-marketing/ directory. Boundary markers: The instructions do not define delimiters or provide warnings to the agent to ignore potential instructions embedded within these loaded files. Capability inventory: The skill's primary capabilities are limited to text generation, narrative assessment, and communication planning; it does not appear to have file-writing or network-access capabilities. Sanitization: There is no evidence of schema validation or content sanitization for the data retrieved from the local filesystem.

crisis-response