The Agent Skills Directory

[SAFE]: The skill provides structural guidance for website planning and does not contain any executable code, network requests, or obfuscated content.- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill is instructed to read local context files like .agents/product-marketing.md or .claude/product-marketing.md. This is a standard feature for AI agents to understand business context and does not involve the exfiltration of sensitive system credentials or unauthorized user data.- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied information and project context files. While it represents a surface for indirect prompt injection, it is assessed as safe because the skill lacks dangerous tools or capabilities. • Ingestion points: .agents/product-marketing.md, .claude/product-marketing.md, and product-marketing-context.md as specified in SKILL.md. • Boundary markers: Absent. • Capability inventory: No file system writing, network operations, or shell command execution capabilities identified across any script or configuration file. • Sanitization: Absent.

site-architecture