The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied marketing scenarios alongside historical brand performance data, which presents a surface for indirect prompt injection. • Ingestion points: User-provided 'Scenarios to compare' and local configuration files in ~/.claude-marketing/brands/ (SKILL.md). • Boundary markers: No explicit delimiters or safety instructions are defined to separate user-provided content from the execution context. • Capability inventory: The skill executes a local script revenue-simulator.py via the marketing-scientist agent (SKILL.md). • Sanitization: No sanitization or validation of user-provided data is specified before it is passed to the simulation tool.
[COMMAND_EXECUTION]: The skill invokes an external script to perform its primary function. • Evidence: The skill explicitly runs revenue-simulator.py to calculate scenario outcomes (SKILL.md).

what-if