edge-concept-synthesizer

---
name: edge-concept-synthesizer
description: Use when abstracting raw detector tickets and hints into reusable edge concepts with thesis and invalidation logic before strategy drafting or export.
category: Strategy R&D
version: 1.0.0
tools: []
---

# Edge Concept Synthesizer

## Overview

Create an abstraction layer between detection and strategy implementation. This skill clusters ticket evidence, summarizes recurring conditions, and produces `edge_concepts.yaml` with an explicit thesis and invalidation logic per concept.

## When to Use

- You have many raw tickets and need mechanism-level structure.
- You want to avoid direct ticket-to-strategy overfitting.
- You need concept-level review before strategy drafting.

## Inputs

- Ticket YAML from detector output (exportable and research-only tickets).
- Optional `hints.yaml` for additional context matching.

## Output

`edge_concepts.yaml` containing, per concept: cluster membership, support statistics, an abstract thesis, invalidation signals, and an export-readiness flag.

## Methodology

1. Collect ticket YAML from the auto-detection output.
2. Optionally bring in `hints.yaml` for context matching.
3. Cluster tickets that share the same underlying hypothesis and overlapping conditions, then synthesize each cluster into one concept with a thesis and invalidation criteria.
4. Deduplicate: merge same-hypothesis concepts whose condition overlap exceeds the configured threshold (default 0.75; tune with an overlap threshold, or disable dedup entirely).
5. Require a minimum ticket-support count per concept (default 2) before promoting it.
6. Optionally promote LLM/human hints directly into concepts, capped by a synthetic-to-real ticket ratio (e.g. 1.5x) so synthetic ideas cannot swamp evidence-backed ones.
7. Review concepts and promote only high-support concepts into strategy drafting.