The Agent Skills Directory

[SAFE]: The skill operates within a restricted directory (brand/) and only interacts with standard project files related to marketing and positioning. It does not attempt to access sensitive system locations, environment variables, or user credentials.
[SAFE]: Network activity is restricted to information retrieval through defined platform tools such as last30days, WebSearch, and Exa MCP. Research queries are constructed for legitimate market analysis targeting well-known public domains and community platforms (e.g., TechCrunch, Reddit, GitHub).
[SAFE]: The skill incorporates robust data validation and safety mechanisms. It includes a mandatory user review phase (Phase 4: Validate with User) for all research findings and implements a 'Claims Blacklist' feature that acts as a hard gate for downstream content generation, effectively mitigating risks associated with processing potentially inaccurate external data.
[SAFE]: Defensive programming practices are explicitly mentioned in the safety rules, including the use of validatePathInput() for file paths and rejectControlChars() for text content to prevent path traversal and control character injection.

landscape-scan