The Agent Skills Directory

[SAFE]: The skill implements a robust task-decomposition framework with no execution of external code, sensitive data access, or network exfiltration patterns detected. All references are local markdown files containing guidance and templates.
[SAFE]: The skill instructions include clear 'route-out' logic to prevent the agent from overstepping its intended role into estimation, board management, or project execution without appropriate tool authorization.
[SAFE]: Domain-specific guidance (Game Dev, Marketing, Fullstack) is provided to ensure planning packets are relevant to the user context while maintaining least-privilege principles by only using allowed tools (Bash, Read, Write, Edit, Glob, Grep) for local repository management.

task-planning