The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documentation suggests running a local Python script scripts/check_prose.py to validate Markdown files. Analysis of the script shows it is a benign utility that uses the Python standard library (re, json, pathlib) to perform regex-based text analysis and linting. It does not use any dangerous functions like eval(), exec(), or subprocess to execute arbitrary commands.
[EXTERNAL_DOWNLOADS]: The skill contains references to official Google developer documentation and a public repository from Cursor (Anysphere) as sources for its writing guidelines. These are trusted and well-known technology sources and do not represent a security risk.
[DATA_EXPOSURE]: The prose-checking script reads local Markdown files and a skills-lock.json file to determine which files to lint. This is standard behavior for a local development tool and does not involve unauthorized access to sensitive files or credentials.
[NO_CODE]: Besides the included Python script and its associated test file, the skill consists entirely of instructional Markdown content and agent configuration.

agent-writing