The Agent Skills Directory

[PROMPT_INJECTION]: No evidence of prompt injection or bypass attempts was found. The instructions focus entirely on information architecture expertise and site planning workflows.
[DATA_EXFILTRATION]: The skill does not attempt to access sensitive system files or exfiltrate data. It references internal project context files like .agents/product-marketing.md to inform its tasks, which is standard behavior for project-aware agents.
[REMOTE_CODE_EXECUTION]: There are no commands that download or execute remote scripts. The skill consists of markdown templates and conceptual guidance.
[OBFUSCATION]: No obfuscated code, encoded strings, or hidden content patterns were detected across any of the analyzed files.
[INDIRECT_PROMPT_INJECTION]: The skill has a low-risk profile for indirect injection. While it ingests user-provided information and context files, it lacks dangerous capabilities like file writing or network operations that could be exploited through malicious data.

site-architecture