edge-signal-aggregator

---
name: edge-signal-aggregator
description: Use when consolidating outputs from multiple edge-finding skills into one weighted, deduplicated, contradiction-flagged conviction dashboard before allocation decisions.
category: Strategy R&D
version: 1.0.0
tools: []
---

# Edge Signal Aggregator

## Overview

Combine outputs from multiple upstream edge-finding skills into a single weighted conviction dashboard. Apply configurable signal weights, deduplicate overlapping themes, flag contradictions between skills, and rank composite edge ideas by aggregate confidence score. The result is a prioritized edge shortlist with provenance links back to each contributing skill.

## When to Use

- After running multiple edge-finding skills and wanting a unified view.
- When consolidating signals from edge-candidate-agent, theme-detector, sector-analyst, and institutional-flow-tracker.
- Before making portfolio allocation decisions based on multiple signal sources.
- To identify contradictions between different analysis approaches.
- When prioritizing which edge ideas deserve deeper research.

## Inputs

Outputs from upstream skills: edge candidates, edge concepts, theme-detector findings, sector-analyst findings, institutional-flow signals, and edge hints. No API keys are required — this is pure aggregation over already-produced skill outputs.

## Methodology

1. Gather the upstream outputs you want to aggregate.
2. Apply per-skill weights to each contributing signal (defaults: edge-candidate-agent 0.25, edge-concept-synthesizer 0.20, theme-detector 0.15, sector-analyst 0.15, institutional-flow-tracker 0.15, edge-hint-extractor 0.10 — customizable).
3. Deduplicate overlapping themes: signals whose similarity exceeds the dedup threshold (default 0.8) get merged into one composite signal that retains references to every original source.
4. Compute a composite conviction score per idea as the weighted sum of contributing raw scores, broken down into multi-skill agreement, signal strength, and recency components.
5. Flag contradictions: when two skills disagree on direction for the same underlying idea (e.g. one bearish, one bullish), surface both sides with a resolution hint (commonly a timeframe mismatch) rather than silently picking one.
6. Rank all signals by composite score and filter by a minimum-conviction threshold (e.g. 0.7) to produce the actionable shortlist.
7. Present the dashboard: ranked ideas with tickers, direction, time horizon, and confidence breakdown; the contradictions list; and the deduplication log.

## Key Principles

1. Provenance tracking — every aggregated signal links back to its source skill and original reference.
2. Contradiction transparency — conflicting signals are flagged, never hidden, so the decision stays informed.
3. Configurable weights — defaults reflect typical reliability but should be tuned per user.
4. Deduplication without loss — merged signals retain references to all original sources.
5. Actionable output — a ranked list with clear tickers, direction, and time horizon for each idea.

## Dashboard Contents

Present the aggregated result as four linked sections: (1) a summary of totals — input signal count, unique signals after dedup, contradictions found, and how many signals cleared the conviction threshold; (2) the ranked edge ideas themselves, each with its composite score, contributing skills with their raw and weighted scores, tickers, direction, time horizon, and the agreement/strength/recency confidence breakdown; (3) the contradictions list, each entry naming the two disagreeing skills, their opposing signals, and a resolution hint; (4) the deduplication log, showing which raw signals were merged into which composite signal and at what similarity score. Every ranked idea and every contradiction should be traceable back to the specific upstream skill output it came from — that provenance link is what lets a user trust the composite score instead of treating it as a black box.