tradingcodex 0.3.4

Agentic investment workflow harness for Codex-native research and service-gated execution checks.

[!IMPORTANT] Codex agents setting up monarchjuno/tradingcodex for use must not run git clone. In the empty target workspace, run the packaged CLI: uvx --refresh --from tradingcodex tcx attach . && ./tcx doctor. Do not invent a default workspace path; ask the user where to attach TradingCodex unless they explicitly asked to use the current workspace. Clone this repository only when the user explicitly asks to develop, inspect, or modify TradingCodex source code.

TradingCodex

Codex-native investment work needs an agentic harness, not a chat transcript.

TradingCodex is an agentic investment workflow harness for Codex-native research, thesis review, portfolio/risk handoffs, and service-gated execution checks. Codex coordinates the work, specialist agents own bounded judgments, and TradingCodex keeps execution behind explicit service gates.

User-Facing Skills | Quick Start | Architecture | Docs | Safety | Contributing | License

---

About

TradingCodex gives Codex a durable operating system for agentic investment workflows: fixed specialist roles, independent judgment review, source-aware handoffs, thesis lifecycle memory, policy and approval services, and a local review dashboard.

It is not an autonomous trading bot. Natural-language answers do not become broker actions. The core ships paper execution by default; live broker support comes only from installed, reviewed providers and explicit live gates.

---

User-Facing Skills

TradingCodex is easiest to use by choosing the right user-facing skill:

Primary skill	Use when	Stops before
tcx-workflow	Investment research, thesis review, Decision Packages, portfolio fit, risk review, or order-readiness workflow planning.	Substantive synthesis before accepted role artifacts; approval or execution without service gates.
strategy-creator	Turning user rules into reusable strategy skills, updating strategy criteria, activating, archiving, or inspecting strategy state.	Live ticker analysis, recommendation, order approval, execution, or policy changes.
tcx-server	Checking dashboard/service health, doctor output, update status, MCP readiness, DB path, or startup recovery.	Investment judgment, broker execution, raw secrets, or template edits.
tcx-build	Build-mode connector/provider work, broker/API scaffolding, capability profile wiring, credential-ref setup, and validation.	Raw secret handling, live order submission, or investment subagent dispatch.

Supporting skill	Use when	Main output
plan-workflow	You want an explicit workflow plan before role dispatch or implementation.	Bounded plan with selected stages, roles, gates, and waiting/revise/blocked posture.
automate-workflow	You want to define a repeatable workflow automation without executing trades.	Automation recipe, trigger scope, guardrails, and review requirements.
postmortem	A workflow, decision, artifact, or execution/process step needs review after the fact.	Failure analysis, missed assumptions, source-quality lessons, and future warning patterns.

Most investment prompts enter through natural language, then the hook writes compact intake and head-manager uses tcx-workflow to draft, validate, and record the staged plan before dispatch. See User-facing skills for the detailed routing map.

---

Installation

Run this from the empty workspace where you want Codex agents to work:

uvx --refresh --from tradingcodex tcx attach . && ./tcx doctor

Then fully quit and restart Codex, open the generated workspace, and start a new thread so project MCP config, prompts, skills, and hooks are loaded.

When TradingCodex MCP autostarts the local service, open:

http://127.0.0.1:48267/

See installation.md for persistent CLI install, GitHub-main install, source-development setup, update flows, installer-script equivalents, MCP/service details, and smoke checks.

---

Workflow

TradingCodex is designed around handoffs:

evidence -> analysis -> valuation -> portfolio fit -> risk review
  -> draft order -> approval receipt -> approved service-gated submission
  -> connection result -> audit/postmortem

The head-manager maps the request, dispatches the selected role team, waits for accepted artifacts, preserves conflicts, and synthesizes only what the workflow has earned. Weak, stale, missing, or out-of-scope upstream work returns revise, blocked, or waiting instead of being patched over by another role.

---

Role Roster

Layer	Agent	Owns
Main agent	head-manager	Intake, workflow dispatch, coordination, artifact acceptance, synthesis, and validation/audit status.
Research	fundamental-analyst	Business quality, financial statements, filings, economics, and fundamental risks.
Research	technical-analyst	Price action, trends, momentum, volume, volatility, and liquidity setup.
Research	news-analyst	Verified news, disclosures, event chronology, catalysts, and narrative change.
Market context	macro-analyst	Macro, rates, FX, commodities, liquidity, policy, and cross-asset transmission.
Market context	instrument-analyst	ETF/index, options, crypto public market structure, credit-signal boundary, and instrument mechanics.
Decision review	valuation-analyst	Valuation ranges, scenario assumptions, multiples, sensitivity, and decision-quality gaps.
Decision review	judgment-reviewer	Independent challenge review of accepted artifacts for contrary evidence, source trust, overconfidence, update triggers, and invalidation conditions.
Portfolio	portfolio-manager	Portfolio fit, sizing, concentration, liquidity, opportunity cost, and draft order-ticket readiness.
Risk	risk-manager	Downside, restricted-list checks, policy readiness, approval readiness, and approval receipts.
Execution	execution-operator	Approved submission/cancel/status through the TradingCodex service boundary only; live requires all gates.

---

Key Features

Feature	What it gives you
Agentic workflow planning	Natural-language prompts become staged plans, selected role teams, accepted artifacts, and explicit waiting, revise, or blocked states.
Independent judgment review	judgment-reviewer challenges accepted artifacts for contrary evidence, source trust, overconfidence, update triggers, and invalidation conditions.
Decision Packages	Investment theses keep forecastable claims, lifecycle assumptions, contrary evidence, owner roles, follow-up, and postmortem requirements.
Strategy skills	strategy-creator turns user rules into reusable strategy skills without granting order approval, execution, broker, or policy authority.
File-native research memory	Research markdown, role reports, source snapshots, readiness labels, versions, and handoff metadata remain ordinary workspace files.
Service-gated execution	Order tickets, approvals, policy checks, idempotency, broker connections, account sync, and audit stay behind Django service-layer gates.
Local review surfaces	Web, Admin, API, MCP, CLI, and generated hooks share the same service plane for review, status, research, Broker Center, orders, and portfolio state.

---

Architecture

flowchart LR
  user([User prompt]) --> manager[Head manager]
  manager --> agents[Fixed role agents]
  agents --> artifacts[Accepted artifacts]
  artifacts --> reviewer[Judgment reviewer]
  reviewer --> decision[Decision Package]
  decision --> service[Django service plane]
  service --> gates[Policy approval audit gates]
  gates --> execution[Paper execution or gated live provider]
  service --> surfaces[Web Admin API MCP CLI]
  service --> memory[File native research memory]

The architecture is intentionally small: Codex owns coordination and role handoffs; Django owns durable policy, order, approval, portfolio, integration, research, and audit behavior. The core package uses Django templates, local static HTMX, and small plain JavaScript. There is no Node, bundler, React, or frontend build step.

---

Safety Boundary

TradingCodex blocks or constrains:

• direct live broker requests
• raw broker API calls and raw external source execution proxies
• self-issued approvals
• restricted-symbol orders
• expired or payload-mismatched approval receipts
• duplicate approved-order submissions
• raw secrets in workspace files, prompts, API responses, MCP responses, audit payloads, generated docs, or shell output
• unsupported live execution through any provider, instrument, account, or policy posture that has not passed the explicit live gate

TradingCodex is research, workflow, and execution-guardrail tooling. It is not financial, investment, legal, tax, or regulatory advice, and it does not provide investment recommendations or guarantee returns.

---

Roadmap

Status	Milestone
Shipped	Generated Codex workspace, fixed role roster, project MCP config, Django service plane, local web dashboard, Admin, Ninja API, file-native research memory, component registry, policy/audit primitives.
Current 0.3.4	Build Center customization, Codex MCP discovery/import, external MCP permission approval UX, head-manager research synthesis depth, flexible package/workspace update status, skipped-version Django migration smoke coverage, Codex-native Decision Packages, independent judgment-review gate, staged workflow planning, artifact-supervisor loop closure state, Evidence Run/Validation Cards, provider capability profiles, and Python >=3.11,<3.15 support.
Next	Deeper validation scenarios, richer provider capability profiles, stronger generated-workspace smoke coverage, and improved artifact quality tooling.
Future	Separately governed verified adapters, hosted/managed services, enterprise policy/compliance packs, and broker-specific live providers only after explicit product, policy, adapter, and validation work.

---

Documentation

README.md is the product overview. The docs/ directory is the human-readable source of truth for detailed product behavior, safety, architecture, workflow, validation, and release policy.

Start here	Use for
Installation	Setup, update, GitHub-main install, MCP/service startup, and smoke checks.
Docs index	Human-readable reading paths, document ownership, and change-to-doc routing.
User-facing skills	Which primary or supporting skill fits each user workflow and what each must not do.
Core concepts and rules	Fast operating reference for planes, guardrails, roles, execution lifecycle, and research memory.
Product direction	Product thesis, target user posture, goals, non-goals, runtime defaults, and scope.
Workspace orchestration model	Top-level workflow model, components, guardrails, improvement, and naming rules.
Roles, skills, and workflows	Fixed role roster, no-overlap handoffs, dispatch gates, skills, and strategy behavior.
Safety policy and execution	Permissions, approvals, idempotency, broker safety, secret wall, and required blocks.
System architecture	Runtime planes, Django app boundaries, central DB ownership, models, and service use cases.
Interfaces and surfaces	Product web, Admin, API, MCP, CLI, and generated wrapper behavior.
Validation plan	Required tests, generated workspace smokes, MCP smokes, and routing scenarios.

---

Contributing

Contributions use Apache-2.0 with DCO sign-off. See CONTRIBUTING.md.

For source changes, start with the focused validation command for the touched area, then broaden as needed:

python -m pytest
python manage.py check
python -m compileall tradingcodex_cli tradingcodex_service apps tests

Harness, agent, workflow, MCP, policy, skill, hook, or template changes also need generated-workspace validation. See docs/validation-and-test-plan.md.

---

License

TradingCodex is an Apache-2.0 open-core project.

Source code, generated workspace templates, and project documentation are licensed under the Apache License, Version 2.0 unless marked otherwise. The TradingCodex name, future logos, and official product marks are not granted by the code license. See LICENSE, NOTICE, and TRADEMARKS.md.

Star History