The Agent Skills Directory

[SAFE]: The skill utilizes the Read tool to access user-provided manuscript files and the Write tool to save review reports locally. These actions are transparently documented and consistent with the skill's primary function of academic paper review.
[SAFE]: All referenced files, such as expression patterns and journal templates, are loaded from local paths within the references/ directory. No external network requests, remote code downloads, or third-party dependency installations were found.
[SAFE]: While the skill processes untrusted user input (manuscripts), which technically creates an indirect prompt injection surface, the risk is mitigated by a strict, locked output structure and a workflow that uses the Write tool with specific, predictable file names. There are no instructions that attempt to bypass AI safety guardrails or extract system configuration.

reviewer-simulation-skill