- Home
- Skills
- Prompt & Skill Engineering
- knowledge base creator pro
knowledge base creator pro
Generates structured Markdown knowledge bases and prompt sequences for AI coding agents.
$29
Works with the AI Agents you already use
knowledge base creator pro
Example session with this skill installed
I want to build a SaaS for fitness tracking using Claude Code and Supabase. Generate the knowledge base for the MVP.
- Read your context and instructions
- Compiled the knowledge base creator
I have analyzed your fitness tracking SaaS idea for Claude Code and Supabase. I have generated the core documentation files including the master plan, architecture, and a 12-prompt sequence to build the MVP.
Knowledge Base for FitnessSaaS
| File | Status | Key Content |
|---|---|---|
00-permanent-rules.md | Created | Rules for TypeScript, Supabase RLS, and modular file structure. |
01-master-plan.md | Created | MVP Focus: Workout logging and progress charts. |
02-design-system.md | Created | High-contrast "Active" theme: #00FF00, #121212. |
03-prompts-claude-code.md | Created | 12 prompts from schema setup to Vercel deployment. |
05-decision-log.md | Created | Why Supabase was chosen over local SQLite. |
06-tasks.md | Created | 24 tasks with "done when" criteria and QA checks. |
07-architecture.md | Created | Schema for profiles, workouts, and exercises. |
10-integrations.md | Created | Supabase Auth and Environment Variable inventory. |
Next steps
- Review
01-master-plan.mdto confirm the workout logging logic. - Check
07-architecture.mdfor theexercisestable fields marked[to verify]. - Open Claude Code and feed it the first prompt from
03-prompts-claude-code.md.
Connects securely to your tools. The creator never sees your data.
About this skill
Coding agents lose focus when they rebuild context at every prompt. This skill builds a durable knowledge base of Markdown files that define rules, architecture, and task sequences to keep your agent on track.
What it does
- Standardized Documentation generates a consistent set of 11 Markdown files including master plans, design systems, and decision logs.
- Agent-Specific Prompting drafts step-by-step prompt sequences tailored for Claude Code, Codex, Lovable, or Replit.
- Integration Orchestration plans modules for auth, payments, transactional email, and storage with environment variable inventories.
- Codebase Auditing reads existing repositories to identify contradictions between code and documentation, reporting discrepancies with severity levels.
- Runtime AI Context crafts system prompts for server-side AI functions, including domain constraints and output style.
How it works
- Input Idea or Repo by describing a new project or pointing the agent at an existing codebase for an audit.
- Specific Interview answers one multiple-choice question per message to fill gaps in the tool, database, and stack requirements.
- Drafting Phase generates the full file set in a
docs/folder or tool-specific directory using provided templates. - Cross-Check Validation runs a verification pass to ensure the data model matches the API contracts and every MVP feature has a corresponding task.
Frameworks & tools
Works with Claude Code, Cursor (Codex), Lovable, and Replit. Supports stacks involving Supabase, PostgreSQL, SQLite, and providers for Auth, Stripe, and AWS/Vercel.
Why this beats prompting it yourself
Manual prompting often leads to "context drift" where the agent forgets architectural constraints or previous decisions. This skill enforces a fixed, version-controlled reference that the agent must reread, reducing hallucinations and wasted tokens.
Use cases
- Scaffold a new MVP from a two-line idea with a full task list and prompt sequence.
- Document a "messy" repository built by an agent to regain control of the architecture.
- Add complex integrations like Stripe or OAuth to an existing agent-led project.
- Generate system prompts for your app's internal AI features based on the project domain.
Known limitations
Does not write application code. Limited to generating the documentation and plan that an agent follows. Requires manual verification of provider-specific API details marked with [to verify].
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Prompt & Skill Engineering.
Does not write application code. Manual verification required for provider-specific facts. Best results require providing a specific tool (e.g., Claude Code, Lovable).
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 day ago
- Software
GitClaude Codecodex