Skip to content

trading.plan.ai

The home of the Trading-Agency — a Hermes-orchestrated, multi-agent US-equity stack, long-only at its root, leverage-extended in the lab, and paper-first. These docs are the spec, the build plan, and the standing orders. The trading system itself is not implemented yet; this is the blueprint for building it.

Read the Build Plan Master Playbook

Master Playbook

The canonical 11-role specification, trade lifecycle, and cross-cutting safety principles. Read it.

Build Plan

Council-vetted, phased plan with safety-first sequencing. Read it.

finance-skills

What to adopt (stripped) — and what to reject. Read it.

Agent guidance

AGENTS.md — standing orders, repo overview, architecture, and the mandate.

The place_order boundary is the central design line. Left of it, language models may propose (screening, debate, sentiment). Right of it, everything is deterministic, unit-tested code (sizing, execution, exits, compliance, the risk/leverage engine). Safety is code, not prompt.

Leverage & Multipliers

The mandate extension — a deterministic, hard-capped risk/exposure engine. Read it.

Cockpit & Frontend

Public landing + private operator cockpit, observing and approving only. Read it.