startup-competitors

# Startup Competitors

Deep competitive intelligence that goes beyond surface-level profiles. Produces actionable battle cards, pricing landscape analysis, and strategic vulnerability mapping using real web data.

## How It Works

```
INTAKE → RESEARCH (3 sequential waves) → SYNTHESIS → BATTLE CARDS
```

The process is focused: understand the product, research competitors deeply across 3 dimensions, synthesize findings, and produce actionable output. Typical runtime: 15-25 minutes in Claude Code (parallel agents), 30-45 minutes in Claude.ai (sequential).

### Language

Default output language is **English**. If the user writes in another language or explicitly requests one, use that language for all outputs instead.

---

## Phase 0: Resume Check

Before anything else, check if a `PROGRESS.md` created by this skill exists in the working directory or a project subdirectory (the skill name field says `startup-competitors`). If it does, read it and resume from the last incomplete phase. Tell the user: "I found progress from a previous session. You completed [phases]. Picking up from [next phase]."

If no progress file exists — or the one found belongs to a different skill — start from Phase 1.

---

## Phase 1: Intake

Short and focused — 1-2 rounds of questions, not an extended interview. The goal is just enough context to run targeted research.

### Check for Prior startup-design Work

Before asking questions, check if a `startup-design` session has already been completed for this project. Look for these files in the working directory or subdirectories:

- `01-discovery/competitor-landscape.md` — competitor profiles and analysis
- `01-discovery/market-analysis.md` — market size, trends, regulatory
- `01-discovery/target-audience.md` — customer personas, pain points
- `00-intake/brief.md` — product description and context

If these files exist, read them and use the data as a head start:
- Extract the product description, target market, and known competitors from the brief
- Use the competitor list from `competitor-landscape.md` as the starting point for deeper analysis (startup-design profiles 5-8 competitors at surface level — this skill goes much deeper on each)
- Pull market size and trends from `market-analysis.md` to contextualize the competitive landscape
- Use customer pain points from `target-audience.md` to focus the sentiment mining on what matters most

Tell the user: "I found data from a previous startup-design session. I'll use it as a starting point and go deeper on the competitive analysis."

Skip the intake interview entirely if the startup-design files provide enough context. Go straight to research.

### What to Ask (if no prior data exists)

**Round 1 — The basics:**
- What's your product/idea? (one sentence is fine)
- What problem does it solve and for whom?
- What market/category are you in?
- Do you know any competitors already? (names, URLs)

**Round 2 — Sharpening (only if needed):**
- What geography/market are you targeting?
- What's your pricing model or range?
- What do you consider your key differentiator?

Don't over-interview. If the user gives a clear description upfront, skip straight to research. The competitive analysis itself will surface what matters.

### Output

Save to `{project-name}/intake.md` — a brief summary of the product, market, and known competitors. If built on startup-design data, note the source files used. The project name should be derived from the product/market (kebab-case, e.g., `ai-email-assistant`).

Create `{project-name}/PROGRESS.md` with: project name, skill name (`startup-competitors`), start date, language, research mode (Live / Knowledge-Based), and a phase checklist. Update it after each phase completes. If PROGRESS.md already exists from a previous session, resume from the last incomplete phase.

---

## Phase 1.5: Research Depth Assessment

After intake, assess market complexity and present the Research Depth recommendation to the user.

> **Reference:** Read `references/research-scaling.md` for the complexity scoring matrix, tier definitions, wave configurations, and the user communication template.

### Process

1. Score three factors from the intake: market breadth (1-3), known competitors (1-3), geographic scope (1-3)
2. Sum the scores (range 3-9) and map to a tier: Light (3-4), Standard (5-7), Deep (8-9)
3. Present the Research Depth table to the user (see `research-scaling.md` for the exact template)
4. Wait for user response: **light**, **deep**, or **ok** to accept the recommendation
5. Record the selected tier in PROGRESS.md

The selected tier determines the number of agents per wave and search rounds per agent in Phase 2. See `research-scaling.md` for exact wave configurations per tier.

---

## Phase 2: Research

Three sequential research waves, each attacking the competitive landscape from a different angle — agents within a wave run in parallel. Together they produce a 360-degree view.

### Environment Detection

Check if the `Agent` tool is available:

- **Agent tool available (Claude Code):** Spawn all agents within each wave in parallel. This is faster.
- **Agent tool NOT available (Claude.ai, web):** Execute research sequentially, following the same templates. Same depth, just slower.

### Web Search

This skill requires WebSearch for real data. If WebSearch is unavailable or denied, fall back to **Knowledge-Based Mode**: use training data, mark all findings with **[Knowledge-Based — verify independently]**, and reduce confidence ratings by one level.

> **Reference:** Read `references/research-principles.md` before starting any wave. It defines source quality tiers, cross-referencing rules, and how to handle data gaps.

### Wave 1: Competitor Profiles + Pricing Intelligence

> **Reference:** Read `references/research-wave-1-profiles-pricing.md` for agent templates.

Two agents (or two sequential blocks):

**A1: Competitor Deep-Dives** — Identify and profile 5-8 direct competitors plus 2-3 adjacent soluti