tradermonty/claude-trading-skills
A Claude skills toolkit turns market review, risk checks, trade planning, and journaling into repeatable workflows, with human decision gates instead of automated trading.
The repository offers **six skillsets** and workflow manifests that specify which skills, decision gates, and artifacts to use. A five-skill starter path can work with public CSVs and local files, without paid market-data APIs.
For agent builders, the useful pattern is to make recurring work explicit: link skills into operational workflows, keep risk checks and review steps visible, and journal decisions for later analysis. The toolkit is designed to structure a person’s process, not choose trades for them.
The repository offers **six skillsets** and workflow manifests that specify which skills, decision gates, and artifacts to use. A five-skill starter path can work with public CSVs and local files, without paid market-data APIs. For agent builders, the useful pattern is to make recurring work explicit: link skills into operational workflows, keep risk checks and review steps visible, and journal decisions for later analysis. The toolkit is designed to structure a person’s process, not choose trades for them. It does not place orders or provide signals. Some workflows depend on optional services such as FMP or Alpaca; even the no-paid-API path still needs external data or local inputs, and the project warns that AI analysis and backtests do not guarantee results.
This describes a process-structuring toolkit for human-directed trading work: manifests connect skills, gates, and journal artifacts, while the project explicitly stops short of trade signals or order placement. The public-data starter path makes paid market services optional, though external or local data is still required.