Idea Validation Agents: app idea validation
idea-validation-agents
Skills for validating an app idea: trends, competitors, pricing, retention, acquisition cost and a final verdict scored 0 to 100
- 511 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The skills write files into memory/ and use web search for trends and reviews
- memory/ stores the author's profile: experience, skills and goals from the interview
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/idea-validation-agentsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skills write files into memory/ and use web search for trends and reviews
- memory/ stores the author's profile: experience, skills and goals from the interview
This is third-party code. Review the repository files before installing.
What it does
The repository turns Claude Code, Codex CLI or Cursor into an analyst for app ideas. Fifteen skills sit in skills/, and thin stubs in .claude/skills/, .codex/skills/ and .cursor/rules/ connect them to each client. Four workflows in the workflows/ folder chain the skills: idea generation, validation, a market deep dive and a pivot. Validation runs in nine steps: trends, a competitor map, desire scoring, pricing, distribution channels, retention, acquisition cost, a final score from 0 to 100 and a decision memo. The score applies a multiplicative floor, so one catastrophic metric drags down the whole result. decision_memo.md records the verdict, risks, the riskiest assumption with a two-week experiment under 100 dollars, and kill criteria. Each step's result is saved in the memory/ folder, and a new session continues where the last one stopped. No packages or API keys are needed: clone the repository and open the folder in the client.
Who it is for. Indie app developers and solo founders who are choosing what to build and want to test an idea in ten to fifteen minutes before writing code.
Good fit when
- You have an app idea and want to know before development whether it is worth building: the verdict and kill criteria land in one memo
- You have no idea but you do have skills and an audience: the agent interviews you about your background and returns 7–10 scored ideas
- A market is chosen, for example meditation apps, and you need a review: competitors, TAM, SAM and SOM sizes, acquisition channels
- Validation ended with a pivot verdict and you need two or three pivot options with projected scores
- You need a cheap experiment for the key assumption that can run before any code is written
Not a fit when
- The idea is not a mobile app: the scoring is tied to the App Store, ASO and mobile retention
- You need verified market figures from paid databases: the calculations rest on open signals and modeled estimates
- The product is already in development and you need code changes: the skills assess the idea and write no code
Example request
I want to build a habit tracker for climbers with daily streaks and tips. Is it worth it?Limitations
Launching is described from the cloned repository folder: the routing rules live in CLAUDE.md, AGENTS.md and .cursor/rules/ of that directory.
Scores, prices and retention forecasts are computed by the model from the formulas in the skills. It does not check sales or the behavior of real users.
Prompts for collecting trends from TikTok, Reddit, X and the App Store sit in skills/trend-analysis/prompts/. Without web access for the agent this step comes out incomplete.
How to disable. The skills work only inside the cloned folder: delete the repository folder. Validation results sit in the memory/ folder of the same directory, so save the verdicts you need from decision_memo.md beforehand.
FAQ
How do I run idea-validation-agents in Claude Code?
Clone the repository, open its folder in Claude Code and describe your idea in the chat. The agent activates from CLAUDE.md and picks the scenario itself. Codex reads AGENTS.md, and Cursor loads the rules from .cursor/rules/.
Do I need API keys or package installs?
No, cloning is enough. If you choose Codex, the Codex CLI itself is installed separately, but the repository's skills need no extra packages.
How long does validating an idea take?
Each scenario, whether idea generation, full validation or a market deep dive, is sized at 10–15 minutes.
What is kept between sessions?
The memory/ folder: the author's profile in user_profile.md, market summaries in market_insights/ and, for each idea, an ideas/<slug>/ folder with competitor, pricing and score files and decision_memo.md. A new session reads them and carries on.
How is the final score computed?
With a multiplicative-floor algorithm: one catastrophic metric, retention for example, drags down the whole score. A real startup works the same way, where one weak link breaks the idea.
Discussion
No questions or reviews yet.
Related
SkillsA skill that scores a startup idea on 14 dimensions with web research: scorecard, assumption ledger and HTML report.
Eleven skills that test a business idea before launch: competitors, a panel of 100 buyers, pricing, a financial model and a 30-day plan
An analytical skill that tests whether users will find value in a product idea: scenario, conditions, timing and delivery, with solidity ratings
A skill that takes a beginner from a vague app idea to a validated plan, then guides the build.
A collection of skills for decisions and projects: strategy, market research, thinking, data analysis, docs and slides.
A Claude Code skill that tests a business idea with a four-agent council and rules: build, fix first or kill.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Idea Validation Agents: app idea validation (this card) | no | 511 | ||
| Startup Skill | no | 1,290 | a plugin for market research, positioning and a pitch, with no numeric multiplicative-floor score | |
| Idea Reality MCP | no | not required | 832 | searches existing projects on GitHub, npm and PyPI and returns a 0 to 100 competition score |
README in short
Skills are described as independent blocks: each has its own input and an output file in memory/, and they never call each other. The lead agent sets the order, reading the previous step's result and passing context to the next. MIT license.
README badge
Original SKILL.md
--- name: decision-memo description: Writes a concise, human-readable decision brief summarizing the full validation analysis — including score, verdict, RAT experiment, pre-mortem, and tier-appropriate next actions. The document a founder actually acts on. --- <!-- version: 0.2.0 | outputs: memory/ideas/<slug>/decision_memo.md --> # Skill: decision-memo ## Purpose After all analysis is complete, produce a document the user can actually act on. This is not a report — it is a **decision brief**. It surfaces the most important signals, names the riskiest assumption, and gives a concrete next action calibrated to the founder's tier. A good decision memo makes the reader feel slightly uncomfortable — that means it's honest. ## Input - Idea slug - `memory/ideas/<slug>/scores.json` (required — includes RAT) - `memory/ideas/<slug>/weaknesses.json` (if available) - `memory/ideas/<slug>/pivot_options.json` (if available and verdict is "pivot") - All other available dimension files in `memory/ideas/<slug>/` - `memory/user_profile.md` (for tier-appropriate recommendations) ## Writing Principles The memo must be **scannable in under 2 minutes**. Follow these rules: 1. **No hedging.** "This might work if…" is banned. State the verdict and own it. 2. **Evidence over opinion.** Every strength and risk must cite a specific data point from a dimension file (k-factor, LTV:CAC ratio, D30 retention, WTP range, etc.). 3. **Asymmetric emphasis on risks.** Humans overweight strengths and underweight risks. The memo corrects for this by giving risks more detail than strengths. 4. **One clear next action.** Not three options — one. The alternative path exists only as a contingency. 5. **Respect the founder's tier.** Don't tell a beginner to "optimize your Meta ads funnel." Don't tell a growth-tier founder to "watch some TikTok tutorials." ## Formatting Constraints | Section | Max length | Purpose | |---|---|---| | Verdict line | 1 sentence | Instant signal | | Score + confidence | 1 line | Quantitative anchor | | Validation watermark | 1–2 lines | Trust calibration (only if confidence < high) | | Top 3 Strengths | 1 sentence each, with one data point | What's working | | Top 3 Risks | 2 sentences each: the risk + what happens if ignored | What kills it | | Riskiest Assumption | 3–5 sentences | The one thing to test before building anything | | Pre-mortem | 3 bullet points | Failure imagination exercise | | Recommended Next Step | 2–4 sentences with specifics | What to do this week | | Kill criteria | 1–2 sentences | When to walk away |