tzurot-council-mcp

The tzurot-council-mcp skill acts as a high-level technical advisor designed for complex software development within the Tzurot ecosystem. It is intended for senior developers, architects, and engineers who face non-trivial technical challenges where multiple potential solutions exist or where high-stakes code modifications—such as major refactors exceeding 500 lines—are required. By providing structured consultation, it helps prevent architectural drift, reduces technical debt, and minimizes the risk of regressions in sensitive services like ai-worker or the voice-engine.

• Performs comprehensive code reviews with a focus on behavior preservation, edge cases, and architectural integrity in TypeScript environments.

• Facilitates structured debugging procedures by requiring detailed error context, code snippets, and prior investigation logs to reduce noise and sharpen diagnostic accuracy.

• Offers deep-dive brainstorming for high-risk refactoring plans, helping to identify potential pitfalls in complex systems like personality memory handling or pgvector integrations.

• Provides intelligent model selection recommendations to ensure the most cost-effective and task-appropriate model (e.g., reasoning vs. coding vs. long-context) is used for specific workloads.

• Enforces strict AI model versioning protocols, prioritizing real-time model list discovery to mitigate issues caused by rapid AI provider model drift.

• Designed for complex architectural decisions, major code refactors, and multi-turn analytical problem solving.

• Users must explicitly call the provided procedural tools (e.g., mcp__council__debug, mcp__council__refactor) to trigger the specific reasoning paths.

• Always call mcp__council__list_models before initiating long-running tasks to prevent 404 errors caused by deprecated model identifiers.

• Consult this skill when you find yourself asking, 'Is this necessary?' before deleting code or changing core service abstractions.

• Strictly avoid using this for trivial tasks, documentation lookups, or simple syntax fixes, as it is optimized for high-complexity engineering deliberation.