The Agent Skills Directory

Indirect Prompt Injection (HIGH): The skill is designed to ingest and process untrusted external content (UI code files under **/aceternityui/**/*.*) using its Read capability. Because the skill also possesses Write and Edit tools, it is vulnerable to indirect prompt injection where malicious instructions hidden in the processed code (e.g., in comments) could trick the agent into performing unauthorized file system modifications.
Ingestion points: File reading of **/aceternityui/**/*.* via Read tool.
Boundary markers: Absent. No instructions provided to ignore or escape embedded instructions within the UI code.
Capability inventory: Read, Write, Edit, and shell execution via cat.
Sanitization: Absent. No validation or filtering of the content being reviewed.
Data Exposure (MEDIUM): The skill includes a 'Memory Protocol' section that explicitly instructs the agent to execute cat .claude/context/memory/learnings.md. This targets internal agent state and metadata files which may contain sensitive project information or previous session context.
Command Execution (LOW): The skill uses a bash block to perform a read operation (cat). While the specific command is benign, it establishes a pattern of direct shell execution to retrieve context from specific paths.

aceternity-ui-configuration