Edge Strategy Designer takes the edge_concepts.yaml written by the concept synthesizer and produces concrete strategy drafts. Each concept can yield core, conservative and research-probe variants, built under a risk profile you choose: conservative, balanced or aggressive.
Exit settings such as stop-loss, reward-to-risk, time stop and trailing stop are adjusted for each hypothesis type, and reward-to-risk is held at a floor of 1.5. The files are research drafts and not investment advice.
What You Get
- strategy_drafts/*.yaml with one draft per concept variant
- A run_manifest.json describing the run
- Optional exportable_tickets/*.yaml for downstream export
- Exit parameters tuned per hypothesis type such as breakout, earnings drift and panic reversal
Common Use Cases
- Producing several variants of each reviewed concept in one run
- Generating exportable tickets to hand to the candidate export step
- Comparing the same concepts under conservative and aggressive risk profiles
What Makes It Unique
Exit calibration is tied to the hypothesis type and applied by script, and the reward-to-risk floor stops drafts from failing the later exit-calibration check in the reviewer. It produces core, conservative and research-probe variants side by side so a reviewer sees the same idea at different strictness levels.
Under The Hood
The agent runs one local Python script that reads edge_concepts.yaml and writes draft YAML, a manifest and optional ticket files to directories you name. It needs Python and PyYAML, uses no API keys and, from what the skill describes, makes no web calls. The agent runs shell commands and writes files.