The Agent Skills Directory

[DATA_EXPOSURE_&_EXFILTRATION]: The skill reads from and writes to the local file system. Specifically, it accesses reference materials and schemas within the skill's root directory and writes UX maps to the docs/ux-research/maps/ path. No sensitive file paths (e.g., SSH keys, credentials) are accessed, and no network exfiltration patterns were detected.
[COMMAND_EXECUTION]: The skill dispatches tasks to other internal agent files (persona-builder.md and renderer.md). This is a standard modular design for AI agents and does not involve arbitrary shell command execution or privilege escalation.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input to define persona characteristics and user journey phases. While this constitutes an injection surface, the risk is inherent to the skill's purpose as a mapping tool and is mitigated by the structured JSON generation process. There are no high-privilege capabilities (like network access or system modification) for an injection to exploit.

experience-map