The Agent Skills Directory

[SAFE]: The skill orchestrates a multi-advisor debate by dispatching local sub-agents using the platform's Agent tool. All advisor logic is contained within the .claude/agents/ directory, and the facilitation process is governed by internal protocol documents.
[SAFE]: No network operations, credential access, or external downloads were detected. The skill reads exclusively from local paths within the skill's own structure (references/, assets/, and .claude/agents/).
[SAFE]: The skill implements significant logical guardrails, including 'Debate Protocols' that require agents to present the strongest version of opposing views and 'Authenticity' checks to prevent agents from compromising their designated priorities for convenience.

council