product-research
Skill0 starsChanged 14 days agoApache-2.0View on GitHub
Validate product ideas from Reddit pain points, competitor reviews, and market signals, with a scored list. Use when researching a niche or an app idea.
14 other projects use a version of this.
Adopted by
Most recently taken up by
- darkweb19/skillstheir own version, first seen Nov 8, 2026
- ethanratnofsky/skillstheir own version, first seen Nov 8, 2026
- newkirkjosh/skillstheir own version, first seen Oct 25, 2026
- liangdabiao/amazon-sorftime-research-MCP-skilltheir own version, first seen Oct 13, 2026
- rolandtolnay/llm-toolkittheir own version, first seen Oct 11, 2026
- veithly/claude-product-workflowtheir own version, first seen Oct 9, 2026
Other versions14 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 14
- liangdabiao/amazon-sorftime-research-MCP-skill★ 1.1ksame name
- AlpacaLabsLLC/skills-for-architects★ 420same name
- mlopscommunity/Coding-Agents-Conference-skills★ 38same name
- Hassaan146/claude-skills★ 18same name
- Ethan-paseya/amazon-sellersprite-research-MCP-skill★ 2same name
- Hussain-Elfalah/Agentic-Skills★ 2same name
- 199-biotechnologies/product-research★ 1same name
- ccgargantua/gars-claude-skills★ 1same name
- mahiunfiltered/claude-code-skills★ 1same name
- newkirkjosh/skills★ 1same name
4 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- AGENTS.mdAGENTS.md
- cloudflare-backendSkill
- composition-rootSkill
- tanstack-frontendSkill
- tech-specSkill
- typescript-standardsSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. 0 are common, 13 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 13 edited versions we read keep these.
13
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Pass 2 (Corroboration): Verify freshness, check existing solutions, read competitor reviews, confirm the pain still exists today
- Distinctive: line 34
Check subscriber counts to prioritize (biggest first)
- Distinctive: line 54
Read 4-5 highest-signal threads.
- Distinctive: line 75
Only after understanding the landscape from Phase 2-3.
- Distinctive: line 83
Create a raw list of pain points with thread URLs and freshness tags before moving to Pass 2.
- Distinctive: line 95
Search for the same pain in recent (last 3 months) posts
- Distinctive: line 96
If the pain is old (e.g., PPC was painful in 2019) — check if the same complaints exist in 2025-2026 threads, or if new tools solved it
- Distinctive: line 97
Look for "I switched to X and it fixed it" comments — that means the problem may be solved
- Distinctive: line 138
Check Chrome Web Store for related extensions (install counts = market size proxy)
- Distinctive: line 139
Check ProductHunt for similar launches (comments show reception)
- Distinctive: line 157
If user says "solo dev" or "easy to build" → weight Ease higher (multiply E by 1.5)
- Distinctive: line 158
If user says "complex is fine" → don't penalize Ease
- Distinctive: line 159
If user specifies revenue model → factor willingness-to-pay signals from threads
“Common” is measured: kept by at least 6 in 10 of the 13 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Product Research & Idea Validation
What this skill does
Two-pass research process to find and validate product opportunities in a target niche:
- Pass 1 (Reddit): Raw pain points, complaints, tool requests from real community conversations
- Pass 2 (Corroboration): Verify freshness, check existing solutions, read competitor reviews, confirm the pain still exists today
Inputs
The user provides:
- Niche/audience (e.g., "Amazon FBA sellers", "Shopify store owners", "freelance designers")
- Optional constraints (e.g., "solo dev", "prefer simple to build", "B2B only")
Pass 1: Reddit Community Research
Browser setup
Always use playwright-cli open --headed --persistent. Both flags required — Reddit blocks if either is missing. Headed avoids headless fingerprint detection, persistent preserves login cookies.
Phase 1: Identify communities (1 min)
- Open browser:
playwright-cli open --headed --persistent - Navigate to Reddit, identify top 2-3 subreddits for the niche
- Check subscriber counts to prioritize (biggest first)
Phase 2: Landscape scan (3-5 min per subreddit)
For each subreddit, extract top posts using JS eval (NOT snapshot parsing):
# Navigate to top posts
playwright-cli goto "https://www.reddit.com/r/{subreddit}/top/?t=year"
# Extract titles + metadata in one shot
playwright-cli eval "JSON.stringify([...document.querySelectorAll('a[href*=\"/comments/\"]')].map(a => ({title: a.textContent?.trim()?.slice(0,150), href: a.href})).filter(x => x.title && x.title.length > 20 && !x.title.includes('Skip') && !x.title.includes('Navigate')).reduce((acc, x) => { if (!acc.seen.has(x.href)) { acc.seen.add(x.href); acc.result.push(x); } return acc; }, {seen: new Set(), result: []}).result.slice(0, 25))"Repeat for t=month to get recent trends.
Time bias: prioritize 2026 > 2025. Threads older than 2024 are background context only — never primary evidence.
Phase 3: Deep-dive threads (2-3 min per thread)
Read 4-5 highest-signal threads. Extract comments:
playwright-cli eval "JSON.stringify([...document.querySelectorAll('[id*=\"comment\"] p, article p')].map(p => p.textContent?.trim()).filter(t => t && t.length > 25).slice(0, 35))"What to look for in comments:
- Complaints about existing tools ("X is clunky", "X doesn't do Y")
- Explicit asks ("I wish there was...", "does anyone know a tool that...")
- Workarounds people describe (manual processes = automation opportunity)
- Recurring frustrations (same problem in multiple threads = high confidence)
- Tool recommendations + why they're insufficient
Track freshness for every data point:
- Note the age of each thread/comment (Reddit shows "Xmo ago", "Xy ago")
- Flag anything older than 12 months as potentially stale
- Recent threads (< 3 months) with the same pain = strong signal
- Old threads with no recent equivalents = possibly solved or irrelevant
Phase 4: Targeted search (only if needed, 2 min)
Only after understanding the landscape from Phase 2-3. Use informed queries based on patterns already observed, not generic "wish there was a tool" queries.
playwright-cli goto "https://www.reddit.com/r/{subreddit}/search/?q={informed_query}&sort=relevance&t=all"Phase 5: Compile Pass 1 findings
Create a raw list of pain points with thread URLs and freshness tags before moving to Pass 2.
Pass 2: Corroboration & Validation
For each top 5-7 ideas from Pass 1, verify with external sources. This is what separates "someone complained on Reddit" from "this is a real, current, underserved market."
2A: Freshness check
For each idea, ask: Is this pain point still active TODAY, or was it solved / became irrelevant?
- Search for the same pain in recent (last 3 months) posts
- If the pain is old (e.g., PPC was painful in 2019) — check if the same complaints exist in 2025-2026 threads, or if new tools solved it
- Look for "I switched to X and it fixed it" comments — that means the problem may be solved
2B: Existing solution audit
For each idea, find what already exists:
# Search for existing tools
playwright-cli goto "https://www.google.com/search?q={niche}+{pain_point}+tool+OR+software+OR+app"
# Check review sites
playwright-cli goto "https://www.g2.com/search?query={tool_name}"
playwright-cli goto "https://www.capterra.com/search/?query={tool_name}"What to capture:
- Tool name, pricing, rating
- Top complaints in reviews (1-2 star reviews are gold — they show unmet needs)
- Feature gaps mentioned by reviewers
- "I switched FROM X because..." comments
2C: Competitor weakness analysis
For the top 3 ideas, read 1-star and 2-star reviews of existing solutions:
# Extract review text
playwright-cli eval "JSON.stringify([...document.querySelectorAll('[class*=\"review\"] p, [class*=\"Review\"] p, [data-testid*=\"review\"] p')].map(p => p.textContent?.trim()).filter(t => t && t.length > 30).slice(0, 20))"Key patterns in bad reviews:
- "Too expensive for what it does" → price disruption opportunity
- "Clunky UI / hard to use" → UX opportunity
- "Missing X feature" → feature gap opportunity
- "Support is terrible" → service opportunity
- "Hasn't been updated in years" → abandoned product opportunity
2D: Market size signals
Quick checks for each top idea:
- Google the main existing tool — check their pricing page for plan tiers (indicates market maturity)
- Check Chrome Web Store for related extensions (install counts = market size proxy)
- Check ProductHunt for similar launches (comments show reception)
Scoring Framework
| Factor | 1 | 2 | 3 |
| Pain (P) | Nice-to-have | Saves hours/week | Critical to revenue/survival |
| Confidence (C) | 1-2 mentions | 3-5 threads + upvotes | Recurring theme, multiple sources |
| Ease (E) | Complex orchestration, many integrations | Standard web app, some API work | Simple SaaS, weeks to MVP |
| Promotability (M) | Hard to reach, enterprise sale | Niche communities, word of mouth | Reddit/YouTube/SEO natural fit, viral potential |
| TAM (T) | Small subset of niche | Most active members | Everyone in the niche |
| Freshness (F) | Modifier: 0.5x if only old threads, 1x if mixed, 1.5x if hot right now |
Score = P x C x E x M x T x F (higher = better)
Adjustments per user constraints:
- If user says "solo dev" or "easy to build" → weight Ease higher (multiply E by 1.5)
- If user says "complex is fine" → don't penalize Ease
- If user specifies revenue model → factor willingness-to-pay signals from threads
Output Format
# {Niche} App Ideas — Research Report
**Date:** YYYY-MM-DD
**Pass 1 Sources:** {subreddits with member counts}
**Pass 2 Sources:** {review sites, competitor tools checked}
## Scoring Framework
{table}
## Ranked Ideas
### 1. [Idea Name] — Score: X (P:X C:X E:X M:X T:X F:X)
**Problem:** What the audience is struggling with
**Freshness:** {ACTIVE / STALE / EMERGING} — {one-line justification, e.g., "3 threads in last 2 months, complaints intensifying due to 2026 fee changes"}
**Evidence (Reddit):**
- [Thread title](https://reddit.com/r/.../comments/...) — "key quote" (X upvotes, Y comments, Z ago)
- [Thread title](https://reddit.com/r/.../comments/...) — "key quote" (X upvotes, Z ago)
**Existing solutions & gaps:**
- {Tool} (${price}/mo, {rating} on G2) — Gap: "{what users complain about}"
- {Tool} (${price}/mo) — Gap: "{missing feature}"
**Opportunity angle:** What specifically to build differently
**MVP scope:** What v1 looks like + rough timeline
### 2. ...
## Top 3 Recommendation
{Brief rationale for the top picks given user's constraints}Behavioral Rules
- Add small random jitter between actions:
sleep $((RANDOM % 2 + 1))between navigations - Use JS eval for data extraction, not snapshot parsing — 10x faster and more reliable
- Browse organically first — read top posts to understand the landscape before searching
- Don't use overly literal search queries — people don't write "I wish someone built X." They complain, ask for alternatives, describe workarounds
- Read comments, not just titles — titles are clickbait, comments contain the real pain
- Always track thread age — a pain point from 3 years ago with no recent mentions may be solved
- Corroborate before scoring — Reddit gives signal, Pass 2 gives conviction. Don't rank high on Reddit alone
- Close browser when done —
playwright-cli close - Target 10-15 ideas from Pass 1, detail top 5 with Pass 2 — diminishing returns after ~30 min total per niche
- Always use
--headed --persistent— user wants to watch, session persists - Always include thread URLs — every evidence point must link to the source so user can verify firsthand
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

