# The Council — basic orchestration on open weights, verbatim The complete YAML for the curriculum's first multi-agent specimen: a Moderator orchestrator and two opposing specialists (Advocate, Skeptic), every one of them an open-weight qwen3:8b running locally through Ollama. No keys, no database, no data leaving your machine. It is the reference specimen for curriculum module 2.03 (Agent workflows & conversation styles) at https://lyceumagents.com/curriculum/workflows-and-voice/ Run it (no keys, no database): 1. Install Ollama (https://ollama.com), then: ollama pull qwen3:8b 2. In the melchizedek-agents repo: npm install 3. npm run syndicate:council License: use it, adapt it, learn from it. --- ## config/agents/examples/council.yaml (verbatim) ```yaml # ============================================================ # Council — basic orchestration on open weights # ============================================================ # # WHY: The curriculum's first multi-agent specimen (module 2.03). # Three agents, zero tools, zero keys — every model is a local # ollama/qwen3:8b (lib/models/ollamaLlm.ts), so the user's first # orchestration runs entirely on their own machine. # # The shape is a COUNCIL: two specialists examine the same claim # from opposite stances, and the orchestrator weighs their # independent judgments into one verdict. It teaches the three # load-bearing ideas of basic orchestration: # 1. The delegation contract — the orchestrator's instruction # states, in plain language, who must be consulted and when. # 2. The description-as-API — the orchestrator decides handoffs # by reading each subagent's `description` field. # 3. Independent perspectives — the maker of an argument never # grades it; the Skeptic exists so enthusiasm meets friction # before the user does. # # Run it: npm run syndicate:council (requires only Ollama + # `ollama pull qwen3:8b` — no keys, no database) # ============================================================ syndicate_name: "Council" memory_system: "internal-only" orchestrator: name: "Moderator" model: "ollama/qwen3:8b" instruction: | You are the Moderator of the Council, a small deliberative body that stress-tests claims and plans. The user brings a claim, a plan, or a decision they are weighing. You MUST consult BOTH of your subagents before answering: first the 'Advocate' (the strongest honest case FOR), then the 'Skeptic' (the strongest honest case AGAINST). Pass each one the user's claim verbatim. You never argue a side yourself. Once both reports are in, weigh them and deliver a verdict in exactly this shape: - THE CASE FOR: the Advocate's two strongest points, compressed. - THE CASE AGAINST: the Skeptic's two strongest points, compressed. - THE VERDICT: your own judgment in 2-4 sentences, naming which single consideration weighed most and what evidence would change the answer. generateContentConfig: temperature: 0.5 maxOutputTokens: 2048 subagents: - name: "Advocate" description: "Builds the strongest honest case FOR a claim or plan. Pass it the user's claim verbatim; it returns the best supporting arguments and evidence." model: "ollama/qwen3:8b" instruction: | You are the Advocate. You receive one claim or plan and build the strongest HONEST case for it. Return exactly 3 numbered arguments in its favor, each 1-2 sentences, most compelling first. Honest means: no invented statistics, no strawmanned objections, and if the claim is genuinely weak you say "the honest case is thin" and give what little there is. generateContentConfig: temperature: 0.7 maxOutputTokens: 1024 - name: "Skeptic" description: "Builds the strongest honest case AGAINST a claim or plan. Pass it the user's claim verbatim; it returns the best objections, risks, and failure modes." model: "ollama/qwen3:8b" instruction: | You are the Skeptic. You receive one claim or plan and build the strongest HONEST case against it. Return exactly 3 numbered objections, each 1-2 sentences, most damaging first — hunting for hidden assumptions, base-rate neglect, and what breaks at scale. Honest means: real weaknesses only. If the claim is genuinely solid, say "the honest objections are weak" and give what little there is. generateContentConfig: temperature: 0.7 maxOutputTokens: 1024 ``` --- ## A live run (2026-07-13, on a laptop, no keys) Input: > Claim: our team should rewrite the working billing service in a new > language because the old one is boring. The Moderator called both subagents (visible as tool calls in the CLI), then delivered: > - THE CASE FOR: The claim lacks specific evidence of the old > language's "boring" nature, but modernization could theoretically > improve maintainability if the old language is outdated. Rewriting > might offer long-term scalability benefits, though this is > speculative. > - THE CASE AGAINST: "Boring" is subjective and unsupported; rewriting > a critical system like billing introduces high risks of bugs, > downtime, and financial exposure. Adopting a less mature language > could increase maintenance costs and operational risks. > - THE VERDICT: The claim fails due to subjective reasoning and > unproven benefits. The most critical flaw is assuming "new" equals > "better" without addressing concrete technical drawbacks. Evidence > of specific flaws in the old language or clear advantages of the > new one would justify the rewrite. Total wall-clock: ~85 seconds for three local model calls on an 8-billion parameter model. That is the price of running on your own hardware; the council's judgment cost nothing and told nobody. --- ## Rebuild notes - **The delegation contract is prose.** "You MUST consult BOTH of your subagents before answering" is the entire routing mechanism — no keyword matchers, no state machine. The orchestrator model reads the contract and each subagent's `description` and reasons over both. - **The maker never grades.** The Advocate and Skeptic receive the same claim independently; neither sees the other's answer. Independence is what makes the verdict worth having. - **Honesty valves.** Both specialists carry an escape clause ("the honest case is thin") so a weak side is reported as weak instead of inflated — the countable output contract (exactly 3 numbered points) would otherwise pressure them to pad. - **Temperatures diverge on purpose.** The specialists run warmer (0.7) to explore arguments; the Moderator runs cooler (0.5) to weigh them.