The Agent Skills Directory

[SAFE]: The skill is focused on documentation and design alignment. It reads project-related files (such as those in docs/ or design/ directories) and writes design specifications to the local workspace. These operations are within the expected scope of a design assistant and do not target sensitive system files.
[PROMPT_INJECTION]: The skill acts on potentially untrusted data from project files and user-provided screenshots, which represents a surface for indirect prompt injection. However, since the skill's output is restricted to Markdown documentation and it lacks high-risk capabilities like remote code execution or network exfiltration, the impact is minimal.
Ingestion points: User input, screenshots, and project files (Markdown, prototypes).
Boundary markers: None identified in the instructions.
Capability inventory: Reading local project documents and writing Markdown files to the workspace.
Sanitization: No explicit sanitization or instruction filtering for ingested content is described.

design-spec