- Home
- Skills
- Prompt & Skill Engineering
- Repo-to-LLM Context Packager: Turn Any Codebase into Clean,
More screenshots
Works with the AI Agents you already use
Repo-to-LLM Context Packager: Turn Any Codebase into Clean,
Sanitize codebases for LLM context and establish multi-agent governance rules for enterprise microservices. Perfect for NotebookLM and AI Coding Agents.
$9.99
Repo-to-LLM Context Packager: Turn Any Codebase into Clean,
Example session with this skill installed
Sanitize my Java backend repository located at ./services/order-service and generate the multi-agent governance files for a system named OrderManagement.
- Read your context and instructions
- Compiled the repo-to-llm context packager
I have processed the ./services/order-service directory. The codebase has been sanitized of comments and sensitive strings, and the governance framework has been established.
Sanitization and Governance Report
| Artifact | Path | Description |
|---|---|---|
| Sanitized Context | fontes_consolidados_LM/consolidado_backend_20231027_143005.txt | Unified code context with target/ and .env excluded. |
| Security Audit | fontes_consolidados_LM/exclusoes_seguranca_20231027_143005.log | 4 secrets (API keys) and 12 PII-like strings removed. |
| Global Governance | ./AGENTS.md | Master guide for OrderManagement defining service boundaries. |
| Local Governance | ./services/order-service/AGENTS.md | Local instructions for the order-service microservice. |
| IDE Enforcement | ./.github/copilot-instructions.md | Non-negotiable rules for Layered Architecture and Data Isolation. |
Security Summary
- Files Scanned: 142
- Tokens Reduced: ~35% (via comment stripping and build artifact exclusion)
- Sanitization Events: Found and masked
spring.datasource.passwordandjwt.secret.
Next steps
- Upload the
consolidado_backend_*.txtfile to your NotebookLM source library for domain analysis. - Review the
AGENTS.mdmaster guide to ensure all microservice port mappings are correct. - Open the repository in your IDE to verify that Copilot/Claude Code is respecting the new
.githubinstructions.
Connects securely to your tools. The creator never sees your data.
About this skill
Enterprise codebases are often too bloated or sensitive for effective AI assistance. Token waste from build artifacts and the risk of leaking secrets prevent developers from getting clean, actionable context in tools like NotebookLM, Claude Code, or Cursor. This skill automates codebase sanitization and establishes a multi-agent governance framework to keep your LLM context focused and secure.
What it does
- Codebase Consolidation scans repositories to generate a unified, comment-stripped context file optimized for RAG and NotebookLM.
- Secret Sanitization identifies and filters passwords, connection strings, JWT secrets, and PII using regex-based security profiles.
- Agent Governance Setup generates root and subproject
AGENTS.mdfiles to map microservice architectures, ports, and domain boundaries. - IDE Enforcement configures
.github/copilot-instructions.mdto mandate architectural patterns like Data Isolation and Event-Driven communication. - Context Optimization automatically excludes build directories, binaries, and logs to prevent token overflow and hallucinations.
How it works
- Run the consolidation script with a specific profile (e.g.,
java,python) to produce a sanitized.txtcontext file and an audit log of exclusions. - Execute the setup script to detect project manifests and generate a global
AGENTS.mdmaster guide for the repository. - Apply the 4-phase lifecycle (Entendimento, Design, Codificação, Validação) to transition from domain strategy to tactical code execution.
Frameworks & tools
The skill utilizes Python scripts for scanning and regex-based sanitization. It supports enterprise stacks including Java (Maven/Spring), Node.js, Python, and Go, and integrates with IDE agents like Claude Code, Cursor, and GitHub Copilot.
Why this beats prompting it yourself
Manual sanitization is error-prone and misses deep-nested secrets or massive build folders that blow through token limits. This skill provides a repeatable, script-driven workflow that ensures every AI agent in your stack follows the same non-negotiable architectural boundaries.
Use cases
- Preparing a legacy microservices repository for ingestion into NotebookLM.
- Establishing architectural guardrails for a team of developers using IDE agents.
- Sanitizing sensitive internal code before sharing context with external LLM providers.
- Mapping complex port and service dependencies for a multi-agent developer environment.
Known limitations
Requires Python to run the consolidation and setup scripts. Sanitization relies on regex patterns and may require custom profiles for non-standard secret formats.
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.
Requires Python environment for scripts. Sanitization is regex-based and may not catch all obfuscated secrets. Architecture detection supports Maven, NPM, Poetry, and Go Modules.
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 16 days ago
- Software
Python