---
name: deep-market-intel
description: Autonomous deep competitor and market intelligence research agent. Scrapes, analyzes, and produces structured competitor matrices, pricing breakdowns, and strategic SWOT dossiers. Trigger when asked to analyze competitors, research market landscapes, or perform product teardowns.
title: Deep Market Intel
short_description: Autonomous competitor teardowns, feature & pricing matrices, and strategic SWOT dossiers built with zero paid API keys.
primary_category: Research
additional_categories:
  - Marketing
  - Productivity
tags:
  - competitor-analysis
  - market-research
  - pricing-matrix
  - swot-analysis
  - web-scraping
support_email: hotchows@gmail.com
execution_mode: run_online
execution_time_minutes: 2
llm_model: frontier-default
---

# Deep Market Intel

Autonomous competitor and market-intelligence workflow. Uses only free, open-source Python libraries (`duckduckgo-search`, `httpx`, `trafilatura`, `beautifulsoup4`, `lxml`, `jinja2`). No paid API keys.

Install once from this skill directory. If the system Python is PEP 668-managed, use a local venv (do not pass `--break-system-packages`):

```bash
python3 -m venv .venv
.venv/bin/pip install -r requirements.txt
```

Then invoke the scripts with `.venv/bin/python` instead of bare `python3`.

## Phase 1 — Collect open-web evidence

Execute the researcher from the skill root (or with an explicit path to the script):

```bash
python3 scripts/ddg_researcher.py --target "<TARGET>" --competitors "<COMPETITORS>" --output /tmp/intel_raw.json
```

- `--target` is required.
- `--competitors` is optional and accepts a comma-separated string (example: `Jira,Asana`).
- The script searches four angles (`product overview features`, `pricing plans cost`, `reviews complaints pros cons`, `vs competitors alternatives`) plus pairwise comparison queries when competitors are supplied.
- Output is structured JSON: `target`, `competitors`, `timestamp`, and `sources[]` (`query`, `title`, `url`, `snippet`, `extracted_text`).

## Phase 2 — Synthesize the dossier

Read `/tmp/intel_raw.json` and write an in-depth intelligence dossier covering all five sections:

1. **ICP & Core Moat** — who the product is for and the durable advantage.
2. **Feature & Pricing Comparison Matrix** — target vs named competitors / alternatives.
3. **Real User Sentiment & Friction Points** — reviews, complaints, praise, and switching triggers.
4. **Complete SWOT Matrix** — strengths, weaknesses, opportunities, threats.
5. **3 Actionable Strategic Attack Vectors** — concrete moves a rival or challenger could take.

Ground every claim in extracted sources. Quote or cite URLs. Separate observed evidence from inference.

## Phase 3 — Compile the companion HTML dossier

```bash
python3 scripts/report_compiler.py --input /tmp/intel_raw.json --output-html ./Competitor_Dossier.html
```

Present the absolute output path of `Competitor_Dossier.html` to the user.

## Troubleshooting

- If DuckDuckGo rate-limits, wait and rerun. Do not introduce paid search APIs.
- Empty `extracted_text` is acceptable for blocked or non-HTML pages; still use title, URL, and snippet.
- Never invent pricing, feature, or review claims that are not present in the JSON or fetched pages.