The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The script scripts/install-dependencies.sh clones external repositories from GitHub (AgriciDaniel/claude-seo and blader/humanizer) and automatically executes a shell script named install.sh found within those repositories. This logic is triggered automatically by scripts/postinstall.js during the npm installation process.
[COMMAND_EXECUTION]: Multiple scripts (scripts/postinstall.js, scripts/ship.js, and scripts/install-dependencies.sh) use execSync or direct shell commands to perform operations such as git clone, zip, npm view, and npm install. This provides a significant attack surface for command injection if user inputs are not properly sanitized.
[EXTERNAL_DOWNLOADS]: The skill is configured to fetch and analyze content from user-provided URLs for 'tone detection' and 'research' purposes, utilizing the read_url_content and search_web tools. While these are part of the intended functionality, they represent a data ingestion path from untrusted external sources.
[COMMAND_EXECUTION]: The skills/writer/update.md and skills/writer/update.md files instruct the agent to run cat, npm view, and npm install -g commands, which interact with the local system and the global npm registry.
[DATA_EXPOSURE]: The installation process and the skill's operational logic involve reading from and writing to the user's home directory (~/.claude/skills) and local project directories, potentially exposing or modifying sensitive configuration files.

content-writer