competitor-aggregate — Stage 2 synthesis skill

Reads every file under marketing/competitors/*.md (per-competitor research files produced by /competitor-research) and synthesizes them into a single canonical threat matrix. The output is the locked file that downstream skills (positioning, messaging, landing-page-copy, sales-enablement) read.

This is the Stage 2 synthesis step in the article's research → context loop. Stage 1 produces per-item research; Stage 2 locks the canonical version.

When to use

• Day 3 of Example 1: after running /competitor-research on 3-5 competitors, synthesize them
• Quarterly refresh: re-run after individual competitor dossiers update
• Before refreshing /positioning — positioning reads the aggregate, not per-competitor files

When NOT to use

• When fewer than 2 per-competitor files exist (no aggregation needed yet)
• For deep-diving a single competitor (use /competitor-research)
• For sales battlecards (use /sales-enablement after aggregate is locked)

How it works

1. Inputs: none required (skill auto-discovers all files under marketing/competitors/*.md except aggregate.md itself)
2. Reads: every per-competitor dossier in the folder. Excludes the prior aggregate.md if one exists (we're rebuilding it).
3. Produces an aggregate threat matrix with these sections:

• Threat ranking — ordered table: competitor name | threat level | category | primary axis of competition
• Common positioning patterns — themes shared across competitors (e.g., all 3 lead on "model sophistication")
• Where the category competes — the axes the category implicitly accepts as the playing field
• Where the category doesn't compete — gaps in the collective positioning that your wedge can exploit
• Status-quo alternatives synthesized — single deduplicated list across all competitors
• Per-competitor wedge summary — one row per competitor: "where they lose / your wedge against them"
• Watch list — events across the set that would change the matrix

1. Marks the file with status: locked frontmatter — signals downstream skills that this is the canonical version
2. Writes to marketing/competitors/aggregate.md (overwrites prior canonical; the prior version is preserved in git history)

Invoke

/competitor-aggregate

Or scoped to a subset:

/competitor-aggregate synthesize only 0526-bizible.md and 0526-dreamdata.md

Example output structure

---
status: locked
locked_date: 2026-05-26
sources: [0526-competitor-a.md, 0526-competitor-b.md, 0526-competitor-c.md]
---

# Competitor aggregate — threat matrix (canonical)

## Threat ranking

| Competitor | Threat | Category | Primary axis |
|---|---|---|---|
| Bizible-style | PRIMARY | Multi-touch attribution | Model sophistication |
| Dreamdata-style | PRIMARY | Multi-touch attribution | Salesforce-native depth |
| HockeyStack-style | DIRECT | Marketing attribution | Self-serve mid-market |

## Common positioning patterns
... (3-5 bullet themes)

## Where the category doesn't compete (your wedge)
... (the gap your positioning exploits)

## Per-competitor wedge summary
... (one row per competitor)

Dependencies

• Reads from: every marketing/competitors/MMYY-*.md file (per-competitor dossiers from /competitor-research)
• Writes to: marketing/competitors/aggregate.md (canonical, locked)
• Downstream readers: /positioning, /product-messaging, /landing-page-copy, /sales-enablement

Customization

Add a custom dimension to the threat matrix (e.g., "Mid-market vs enterprise split") by editing the section list in this SKILL.md body. The aggregate adapts to whatever sections you specify.

Where this fits in the Example 1 chain

Day 1-3: /competitor-research × N → per-competitor files
Day 3:   /competitor-aggregate (THIS SKILL) → canonical threat matrix
Day 5:   /icp-research reads aggregate.md
Week 2:  /positioning reads aggregate.md

Refresh cadence

Re-run whenever any per-competitor file updates. Quarterly minimum; fortnightly if you're in a fast-moving category.