The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructs the agent to retrieve a GitHub Personal Access Token from a specific local path at outputs/.claude-credentials/github-pat.txt as part of its deployment workflow.
[COMMAND_EXECUTION]: The script scripts/screenshot.py executes system commands using subprocess.run to invoke browser binaries like chromium or google-chrome for rendering UI previews. Additionally, references/deployment.md contains bash snippets for the agent to execute curl and git operations.
[EXTERNAL_DOWNLOADS]: The skill references external assets from fonts.googleapis.com and fonts.gstatic.com. It also includes instructions for the installation of the playwright Python package and its browser dependencies to support the screenshot-loop functionality.
[DATA_EXFILTRATION]: In references/deployment.md, the skill describes a workflow that reads a local credential file and transmits the token to the GitHub API (api.github.com) via curl to automate repository creation. While targeting a well-known service, this involves the movement of sensitive authentication data.

website-builder