The Agent Skills Directory

[SAFE]: The skill is focused on academic paper polishing and technical editing. Analysis of the instructions and supporting files reveals no evidence of credential harvesting, unauthorized network activity, or persistence mechanisms.
[COMMAND_EXECUTION]: The skill utilizes local Python utility scripts (scripts/ai_score_estimator.py, scripts/validate_polish_output.py, and scripts/tier1_words.py) to provide automated checks on the text. These scripts use standard Python libraries (re, json, pathlib) and are limited to processing user-provided text files to ensure the integrity of LaTeX and mathematical content.
[PROMPT_INJECTION]: The skill manages the risk of indirect prompt injection from untrusted academic documents by implementing 'Hard Rules' that strictly forbid the modification of LaTeX commands, technical claims, or citations. It further mitigates this risk through a validation script that audits the final output for consistency with the original document's technical zones.

sci-polish