The Agent Skills Directory

[SAFE]: The skill is composed entirely of markdown-based instructions, guidelines, and templates. It contains no executable code, scripts, or automated tasks that could pose a security risk.\n- [PROMPT_INJECTION]: No evidence of prompt injection, role-play bypasses, or instructions to override safety guidelines was found. The persona remains consistent as an information architecture expert.\n- [DATA_EXFILTRATION]: While the skill references reading local context files such as '.agents/product-marketing-context.md', it lacks any capability to transmit this data externally. There are no network operations (curl, wget, etc.) or outbound communication patterns present.\n- [REMOTE_CODE_EXECUTION]: The skill does not download, install, or execute external packages or scripts. All reference material is static text or diagram code (Mermaid/ASCII) meant for rendering by the AI interface.\n- [COMMAND_EXECUTION]: No shell commands, subprocess calls, or privileged operations were found. The skill operates within the safe boundaries of the agent's standard conversational capabilities.

site-architecture