The Agent Skills Directory

[SAFE]: The skill implements a storytelling persona named Sophia and uses local project files for configuration and memory management, which is standard for agents maintaining state across sessions. No network operations or sensitive system file accesses were detected.- [PROMPT_INJECTION]: Instructions use persona reinforcement ('do not break character') to maintain the user experience without attempting to bypass safety filters. The skill ingests data from local files like project-context.md and story-preferences.md (Ingestion points), but lacks explicit boundary markers or sanitization logic (Sanitization). However, the capability inventory is limited to narrative strategy and help, posing no significant risk of exploitation.- [DATA_EXFILTRATION]: No network activity or exfiltration patterns were identified. The skill only accesses project-specific configuration and memory files within the _bmad directory, which is consistent with the author's namespace.

bmad-cis-agent-storyteller