The Agent Skills Directory

[SAFE]: The skill consists entirely of markdown instructions and configuration. It does not contain any executable scripts, shell commands, or external code dependencies.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted text data (manuscripts) through its file:read permission. This creates a surface for indirect prompt injection, where instructions hidden within a manuscript could attempt to manipulate the AI's assessment. However, the skill does not have access to sensitive capabilities like network communication or system-level command execution, which significantly mitigates the risk and impact of such an injection.

sensitivity-reader