The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script to manage campaign data.
Evidence: SKILL.md contains instructions to run python campaign-tracker.py --brand {slug} --action list-campaigns to retrieve campaign history.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection through the ingestion of external data files.
Ingestion points: The agent reads brand profiles, guidelines, and restrictions from files located in ~/.claude-marketing/brands/{slug}/.
Boundary markers: There are no explicit delimiters or instructions provided to the LLM to ignore potentially malicious content within these brand profile files.
Capability inventory: The skill can execute subprocesses (campaign-tracker.py) and access the file system.
Sanitization: No sanitization or validation logic is defined for the content of the brand JSON or markdown files before they are loaded into the agent's context.

content-engine