The Agent Skills Directory

[SAFE]: The skill performs legitimate analysis of research artifacts and experiment results. It manages project state using local directories (e.g., memory/, paper/.agent/), which is consistent with standard project memory practices. While the skill utilizes tools such as Bash and WebFetch and processes external files (e.g., .tex, .csv, images), these capabilities are strictly aligned with its documented purpose of figure review and research provenance tracking. No evidence of prompt injection, data exfiltration, or unauthorized command execution was found.

figure-results-review