The Agent Skills Directory

[SAFE]: A thorough review of all skill files and referenced guides found no evidence of malicious intent, prompt injection, obfuscation, or unauthorized data access. The skill functions as a structured template and orchestration layer for presentation design.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied topics and context to build narratives. While this constitutes an attack surface for indirect injection, the skill's rigid adherence to specific narrative frameworks (Pyramid Principle, SCQA, SCR) and mandatory slide templates acts as a structural safeguard that mitigates the risk of the agent following instructions embedded in the input data.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill accesses configuration files located in ${CLAUDE_PROJECT_DIR}/.brand-studio/brand-config.md and ${CLAUDE_PLUGIN_DATA}/brand-config.md. These paths are used to retrieve design assets like colors and fonts, which is consistent with the skill's primary purpose of branded deck generation and does not involve sensitive credential harvesting.

deck-builder