The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes local Python scripts located in the ${CLAUDE_PLUGIN_ROOT} directory (sample-size-calculator.py, significance-tester.py, and campaign-tracker.py) to perform statistical calculations and manage brand campaign data.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from brand profiles and guidelines, which constitutes a potential attack surface.
Ingestion points: Reads data from ~/.claude-marketing/brands/{slug}/profile.json and various guideline files such as restrictions.md and messaging.md.
Boundary markers: There are no explicit delimiters or specific instructions provided to the agent to treat the ingested brand data as untrusted or to ignore embedded instructions.
Capability inventory: The skill has the capability to execute shell commands and perform file read operations across the local file system.
Sanitization: No explicit sanitization or validation of the content within the brand profile files is performed before processing.

cro