edge-hint-extractor

1# Edge Hint Extractor
2 
3## Overview
4 
5Convert raw observation signals (`market_summary`, `anomalies`, `news reactions`) into structured edge hints.
6This skill is the first stage in the split workflow: `observe -> abstract -> design -> pipeline`.
7 
8## When to Use
9 
10- You want to turn daily market observations into reusable hint objects.
11- You want LLM-generated ideas constrained by current anomalies/news context.
12- You need a clean `hints.yaml` input for concept synthesis or auto detection.
13 
14## Prerequisites
15 
16- Python 3.9+
17- `PyYAML`
18- Optional inputs from detector run:
19  - `market_summary.json`
20  - `anomalies.json`
21  - `news_reactions.csv` or `news_reactions.json`
22 
23## Output
24 
25- `hints.yaml` containing:
26  - `hints` list
27  - generation metadata
28  - rule/LLM hint counts
29 
30## Workflow
31 
321. Gather observation files (`market_summary`, `anomalies`, optional news reactions).
332. Run `scripts/build_hints.py` to generate deterministic hints.
343. Optionally augment hints with LLM ideas via one of two methods:
35   - a. `--llm-ideas-cmd` — pipe data to an external LLM CLI (subprocess).
36   - b. `--llm-ideas-file PATH` — load pre-written hints from a YAML file (for Claude Code workflows where Claude generates hints itself).
374. Pass `hints.yaml` into concept synthesis or auto detection.
38 
39Note: `--llm-ideas-cmd` and `--llm-ideas-file` are mutually exclusive.
40 
41## Quick Commands
42 
43Rule-based only (default output to `reports/edge_hint_extractor/hints.yaml`):
44 
45```bash
46…