I Built 67 Config Validators and Made $0 — So I Built 88 Agent Skills That Actually Matter

Last month I went all-in on ClawHub, the agent skills marketplace. The plan: find empty niches, build config validators for every popular tool I could think of, price them at $49-99, profit.

I built 67 of them. tsconfig.json, helm values.yaml, pyproject.toml, Docker Compose, nginx, Terraform, CircleCI, you name it. Each one a standalone Python script, zero dependencies. Feed it a config, get back a list of issues.

Total revenue after a month: $0.00.

The embarrassing part

ClawHub does not have a payment mechanism. Everything on the registry is free. My $49-99 pricing tags were just metadata that nobody ever saw.

I could have found this out in five minutes on Day 1. I found it out on Day 67.

Empty niches are not always opportunities

I kept getting excited about zero-competition scores. Terraform validator? No competitors! nginx validator? Nobody!

Nobody was competing because nobody was searching. Developers who use nginx already have nginx -t. TypeScript users have the compiler. ESLint users have ESLint. I spent weeks building tools that compete with built-in features of the tools they validate.

What people actually install

The top ClawHub skills get 10K-35K downloads. They are all agent capabilities: Capability Evolver (35K), WhatsApp integration (16K), browser automation (11K). Not standalone scripts. Integrations that let an AI agent do something it could not do before.

My validators were Python scripts with a SKILL.md stapled on top. They did not use AI reasoning. They did not connect to anything. They just ran a linter.

The pivot: 88 agent capability skills in 4 days

Once I saw the pattern, I changed strategy completely. Instead of config validators, I started building skills that teach agents how to reason about real engineering problems.

Four days later: 88 agent capability skills across 9 categories.

DevOps & Infrastructure (16 skills):
Infrastructure Drift Detector, Config Drift Scanner, Canary Deployment Analyzer, Runbook Automator, Capacity Planner, CI Pipeline Optimizer, Environment Promoter, Post-Deployment Verifier, ArgoCD Deployment Analyzer, Kubernetes Operator Scaffolder, Renovate Config Generator, FinOps Cloud Optimizer, On-Call Schedule Optimizer, Toil Tracker, Service Dependency Mapper, Error Budget Tracker

Security & Compliance (12 skills):
Vulnerability Prioritizer, TLS Configuration Auditor, CSP Policy Generator, Password Policy Auditor, Data Anonymizer, SBOM Generator, Trivy Security Scanner, Branch Protection Auditor, Key Rotation Planner, Certificate Lifecycle Manager, Prompt Injection Tester, Code Review Assistant

Testing & Quality (8 skills):
Mutation Test Runner, Flaky Test Detective, API Contract Tester, Chaos Test Designer, Test Impact Analyzer, LLM Eval Harness, dbt Model Auditor, Airflow DAG Analyzer

Performance & Debugging (8 skills):
N+1 Query Detector, Memory Leak Hunter, Thread Dump Analyzer, Cache Strategy Advisor, Backpressure Analyzer, WebSocket Tester, PostgreSQL Query Optimizer, Bun Workspace Optimizer

Architecture & Planning (10 skills):
Retry Policy Designer, Dead Letter Queue Analyzer, Codebase Migration Planner, ADR Generator, Monorepo Analyzer, Feature Toggle Manager, Shadow Traffic Tester, Dark Launch Controller, Namespace Governance, Autoscaling Policy Designer

Developer Productivity (10 skills):
Codebase Onboarder, Git Changelog Generator, JWT Debugger, OAuth Debugger, Daily Standup Generator, OKR Progress Tracker, Conversation Summarizer, Notion Workflow Automator, Linear Issue Manager, Postmortem Generator

AI/LLM Tools (6 skills):
Hallucination Detector, RAG Chunking Optimizer, Prompt Injection Tester, LLM Eval Harness, Content Repurposer, Prompt Optimizer

Cloud & Kubernetes (8 skills):
Kubernetes Operator Scaffolder, ArgoCD Deployment Analyzer, Trivy Security Scanner, Renovate Config Generator, FinOps Cloud Optimizer, Cluster Upgrade Planner, Cloud Tagging Auditor, DB Replication Monitor

Web & Frontend (4 skills):
Astro Project Analyzer, Site Health Monitor, Sitemap Generator, CORS Scanner

The difference matters. The N+1 Query Detector does not just grep for .find() in a loop. It identifies ORM access patterns, checks if eager loading is possible, estimates the query count per request, and suggests the exact fix depending on the framework. That requires AI reasoning about code structure.

The Hallucination Detector cross-references LLM outputs against source documents, code, and verifiable data. It classifies each finding as fabrication, contradiction, exaggeration, or outdated. The RAG Chunking Optimizer analyzes your document corpus and recommends chunk sizes, splitting methods, and overlap settings tailored to your embedding model and query patterns.

The Vulnerability Prioritizer does not just sort CVEs by CVSS score. It enriches with EPSS exploit probability, checks the CISA KEV catalog, assesses asset criticality. CVSS 9.8 with 0.001 EPSS on an internal service gets deprioritized. CVSS 7.1 with 0.87 EPSS on a public API goes to the top.

The numbers after 4 days

• 155 skills total (67 validators + 88 agent capabilities)
• 111 published on ClawHub, rest publishing via rate-limited drip
• 18 zero-competition categories claimed (infrastructure drift, mutation testing, hallucination detection, RAG chunking, LLM evaluation, ArgoCD, Kubernetes operators, OKR tracking, daily standup generation, and more)
• Revenue: still $0

What I learned

Three things:

Check whether the platform takes money before building 67 things to sell on it. That one is obvious in hindsight.

An empty niche on a marketplace can mean untapped market, but it usually means nobody wants this. The signal is not competition score alone. The signal is what the top skills have in common: they extend what an agent can do. Empty + high demand in the broader dev ecosystem = real opportunity. Empty + nobody asks for this = you are building into the void.

And move fast once you see the pattern. I built 88 skills in four days by systematically scanning niches, filtering for real demand, and targeting the intersection of zero ClawHub competition with high engineering value. Not every skill will get traction. But coverage at this scale means some will find their audience organically.

The revenue will come from off-platform: Lemon Squeezy bundles, setup-as-a-service, consulting driven by the portfolio. ClawHub is distribution, not the storefront.

All skills are free on ClawHub under handle charlie-morrison.