The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill's primary purpose is to provide a framework and checklist for auditing data movement and exfiltration risks in other skills.
[NO_CODE]: The skill contains no executable scripts (Python, JavaScript, or shell), binary files, or automated commands. It functions as a set of human-in-the-loop instructions and reference materials.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to ingest and analyze external data (other agent skills), which is an inherent attack surface for indirect prompt injection, the skill lacks any automated capabilities (like writing to files or making network calls) that could be subverted by malicious input in the skills being reviewed.

data-exfiltration-review