start-council

name

start-council

description

Launch an LLM-Council-Policy-Ideation run — six policy-lens councillors (Urban Planner, Fiscal Officer, Equity Advocate, Legal Analyst, Service Delivery, Innovation Scholar) examine a local-governance policy problem via a single LLM behind different system prompts. Use when the user says "run the council", "deliberate on this policy problem", "/start-council", or provides a problem statement in the LLM-Council-Policy-Ideation repo. Starts the app and surfaces the Chairman's Policy Ideation Report plus optional PDF/podcast digests.

Start a Council (Policy Ideation)

You are operating inside the LLM-Council-Policy-Ideation repo. This variant is tuned for local-governance policy challenges — the user supplies a policy problem statement, six domain-specialist councillors examine it, and the Chairman produces a structured Policy Ideation Report.

Preconditions

1. Repo root must contain backend/main.py, main.py, and start.sh.
2. OPENROUTER_API_KEY must be set. No other keys are required for the base run.
3. uv sync and cd frontend && npm install && cd .. if dependencies are missing.

Step 1 — Capture the policy problem

Ask the user for a problem statement describing the policy challenge. A good statement names:

• The municipality / jurisdiction (or jurisdiction type if anonymised)
• The observable problem (with any numbers or timeframe the user has)
• What prompted the council — a failing programme, a new mandate, a political pressure, a forthcoming decision

If the statement is sprawling, help the user tighten it to a single paragraph. The councillors take the framing literally — a buried assumption will propagate into every lens.

Step 2 — Start the app

./start.sh

Or manually:

# Terminal 1
uv run python -m backend.main
# Terminal 2
cd frontend && npm run dev

Direct the user to http://localhost:5173.

Step 3 — Surface the output

The UI walks through the three stages:

1. Councillor perspectives — six lenses respond independently.
2. Peer review — each councillor ranks the others' (anonymised) responses.
3. Chairman synthesis — the final Policy Ideation Report with restated problem, consolidated idea catalogue, cross-cutting themes, tensions, a shortlist of promising directions, and open questions.

Encourage the user to read at least the Equity Advocate and the Fiscal Officer in full before jumping to the synthesis — those two most often hold the load-bearing objections that the Chairman must reconcile.

Step 4 — Digests

If requested, trigger:

• PDF report — Typst-typeset Policy Ideation Report. Suitable to share with a municipal working group.
• Briefing podcast — LLM-written script → Microsoft Edge TTS audio. Useful for a commuting official.

Step 5 — Iterate

Offer to:

• Re-run with a tightened problem statement.
• Swap or add councillor lenses in backend/config.py for domain-specific policy work (e.g. Public Health Officer for health policy, Climate & Resilience lens for environmental policy).
• Fork the repo for a parallel topic area.

Failure modes

• Missing OPENROUTER_API_KEY → stop.
• Frontend port collision → use an alternate port.
• A councillor returns empty → retry that slot; report if it fails twice.

Out of scope

This variant is for ideation — surfacing and structuring the space of policy options. It is not a decision tool: the Chairman produces a catalogue with tensions named, not a single recommended option. For decision-making between defined options, use LLM-Council-Decide.