The Agent Skills Directory

[SAFE]: The skill is a purely descriptive markdown file that outlines a 'Global Workspace' pattern for agent coordination. It contains no executable scripts (Python, Node.js, or Shell) and no instructions that trigger external tools or network access.
[SAFE]: The 'Security & Guardrails' section within the documentation identifies potential threats like prompt injection (hijacking the coalition manager), data exposure (broadcasting secrets), and denial of service (workspace flooding), and suggests appropriate architectural mitigations.
[SAFE]: No obfuscation, hardcoded credentials, or malicious prompt injection patterns were found. The use of template variables like '{{previous_broadcast}}' is standard for agent instruction templates.

global-workspace