Skip to content

Add Hermes control plane automation foundation#1

Draft
chuckatbcs wants to merge 16 commits into
mainfrom
hermes-control-plane
Draft

Add Hermes control plane automation foundation#1
chuckatbcs wants to merge 16 commits into
mainfrom
hermes-control-plane

Conversation

@chuckatbcs
Copy link
Copy Markdown
Owner

Summary

This PR adds a first foundation layer for running TradingOrg as a Hermes-controlled research and post-game review engine on the home PC.

Added

  • AGENTS.md with repo-wide operating rules for AI agents.
  • docs/HERMES_CONTROL_PLANE.md describing the Hermes/WebUI control-plane workflow.
  • Non-interactive console scripts:
    • tradingagents-analyze-run
    • tradingagents-postgame
    • tradingagents-update-lessons
  • Automation package under tradingagents/automation/.
  • Hermes-safe scripts under scripts/hermes/:
    • run_single_ticker.sh
    • run_daily_watchlist.sh
    • run_postgame_review.sh
  • Signal JSON output for automated runs.
  • Deterministic post-game review JSON generation.
  • memory/lessons.jsonl append workflow.
  • .env.example updates for local Ollama, remote Tailscale Ollama, OpenRouter/free-cloud fallback, and premium fallback model settings.
  • Docker Compose env and volume wiring for reports/ and memory/.

Changed

  • DEFAULT_CONFIG now reads relevant LLM, routing, runtime, and data-vendor defaults from environment variables.
  • OpenAI-compatible provider client now lets explicit base_url override provider defaults, enabling remote Ollama over Tailscale.
  • Ollama model catalog now supports custom model IDs.
  • .gitignore now excludes generated runtime outputs such as reports and JSONL lessons.

Safety boundaries

This PR does not add broker execution. TradingOrg remains a research/signal generator only. Generated signals include paper_only: true and an execution policy stating that the signal must not directly place broker orders.

Testing notes

Not run in this environment. Recommended local checks:

pip install -e .
python -m pytest tests -q
tradingagents-analyze-run --help
tradingagents-postgame --help
tradingagents-update-lessons --help

Suggested smoke test with local/home-PC Ollama:

LLM_PROVIDER=ollama \
LLM_BACKEND_URL=http://home-pc-tailnet-name:11434/v1 \
QUICK_THINK_LLM=qwen3:latest \
DEEP_THINK_LLM=qwen3:latest \
tradingagents-analyze-run \
  --ticker SPY \
  --date 2026-05-19 \
  --provider ollama \
  --backend-url http://home-pc-tailnet-name:11434/v1 \
  --quick-model qwen3:latest \
  --deep-model qwen3:latest \
  --analysts market,news,fundamentals \
  --depth 1 \
  --output-json reports/signals/SPY/2026-05-19/signal.json

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant