council-taxonomy
Council Taxonomy
This reference defines council primitives used by agent-council, deep-council, debate-protocol, and bridge adapters.
Core Principle
Deep insight does not come from model diversity alone. Different runtimes using the same model can produce materially different findings because their planning loop, tool affordances, context packaging, file discovery, browser/repo access, session continuity, and error recovery differ.
Council reports must therefore record all diversity sources:
- Role diversity — different expert perspectives inside one council.
- Model diversity — different model weights/providers.
- Runtime diversity — different agent executors such as Claude Code, Codex, OpenCode, Gemini CLI, or Hermes subagents.
- Toolchain diversity — different available tools, MCP servers, browser/repo access, and shell/file workflows.
- Evidence-channel diversity — different ways of grounding claims: tests, logs, source files, web sources, browser observations, human input.
- Debate-layer diversity — local/intra-council debate and cross-council debate.
Council Types
More from mikeng-io/agent-skills
perplexity
Optional AI-synthesized web search via Perplexity MCP. Returns AI-grounded answers with inline citations rather than raw search results — best for synthesis-heavy questions, current consensus queries, and cross-validation of research findings. Non-blocking — falls back to other search tools if unavailable.
10brave-search
Optional web search enrichment via Brave Search MCP. Provides web, news, and local search with an independent index (no Google dependency). Invokable standalone or called by deep-research, deep-explorer, or any skill needing web search. Non-blocking — falls back to other search tools if unavailable.
9deep-audit
Multi-agent standards and compliance auditing with pass/fail verdicts. Checks against security, accessibility, code standards, regulatory requirements, and performance benchmarks.
9deep-review
Multi-agent quality improvement review with constructive feedback. Provides suggestions for best practices, code quality, alternatives, and performance optimization.
8deep-research
Generic multi-domain research framework with domain-aware scheduling. Model-agnostic and domain-agnostic - perform comprehensive research on any topic.
8context
Analyze conversation context to classify what is being discussed, detect relevant domains from domain-registry, and determine optimal routing (parallel-workflow vs debate-protocol vs deep-council). Can be invoked directly or called by other skills as a pre-flight step.
7