mini-claude-code

🥈Silver

Build custom coding agents with Claude. Operations teams create automated workflows. Integrates with Kode CLI for agent development.

328390Updated 1mo ago

Intermediate30min to implementautomation

Saves ~45 min per use

Quick InstallView Source

git clone https://github.com/shareAI-lab/mini-claude-code.git

Works with:

Claude

Overview

About This Skill

Build custom coding agents with Claude. Operations teams create automated workflows. Integrates with Kode CLI for agent development.

How to Use

[{"step":"Define the agent's purpose and requirements. List the specific task it should automate, the programming language, and any tools/frameworks it must integrate with. Include error handling and logging needs.","tip":"Be specific about edge cases. For example, if validating CSV files, specify how to handle missing values, incorrect data types, or malformed rows."},{"step":"Use the Kode CLI to scaffold the project. Run `kode new [AGENT_NAME] --language [LANGUAGE] --framework [FRAMEWORK]` to generate the initial template and directory structure.","tip":"Review the generated files to understand the default structure. The CLI typically creates a main agent file, configuration files, and a tests directory."},{"step":"Implement the core logic in the generated agent file. Add functions to handle the task, integrate with required tools, and include error handling and logging as specified.","tip":"Start with a minimal viable version of the agent. Test it with a small dataset or subset of requirements before scaling up."},{"step":"Configure the agent's settings. Update the `config.yaml` or equivalent file with parameters like file paths, API keys, or thresholds. Set up logging destinations and levels.","tip":"Use environment variables for sensitive data like API keys. Tools like `python-dotenv` can help manage these in development."},{"step":"Test the agent locally before deploying. Run unit tests and integration tests to ensure it meets the requirements. Use the CLI or a local script to simulate the agent's operation.","tip":"Mock external dependencies (e.g., APIs, databases) during testing to isolate the agent's logic. Tools like `pytest-mock` or `unittest.mock` can be helpful."},{"step":"Deploy the agent to your target environment. Use the Kode CLI or your CI/CD pipeline to build and deploy the agent. Monitor its performance and adjust as needed.","tip":"Set up alerts for errors or failures. For example, use Slack notifications or email alerts to stay informed about issues in production."}]

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/shareAI-lab/mini-claude-code

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Build a custom coding agent using [AGENT_NAME] that automates [SPECIFIC_TASK] in [PROGRAMMING_LANGUAGE]. The agent should integrate with [TOOLS_OR_FRAMEWORKS] and follow these requirements: [LIST_REQUIREMENTS]. Use the Kode CLI to scaffold the project and generate the initial agent template. Include error handling for [COMMON_ERRORS] and logging for [LOGGING_LEVEL].

Example Output

Here’s a custom coding agent named `data-pipeline-validator` built to automate CSV file validation and transformation in Python. The agent integrates with Pandas for data processing, SQLite for metadata tracking, and GitHub Actions for CI/CD. Requirements included: (1) Validate CSV headers against a schema file, (2) Convert data types per column specification, (3) Log validation errors to a Slack channel, and (4) Generate a summary report in Markdown.

The agent was scaffolded using Kode CLI with the command: `kode new data-pipeline-validator --language python --framework pandas`. The initial template included a `validator.py` file with a `CSVValidator` class, a `config.yaml` for settings, and a `tests/` directory with pytest cases. Error handling was implemented for missing files, schema mismatches, and data type conversion failures. Logging was configured to output to both console (INFO level) and Slack (ERROR level) using the `slack_sdk` library.

The agent processes files from an S3 bucket, validates them against a schema stored in GitHub, and moves valid files to a `processed/` folder while logging errors to `#data-errors` in Slack. A summary report is generated for each run, including row counts, validation errors, and processing time. The agent was deployed via GitHub Actions, with the workflow triggering on new CSV uploads to the S3 bucket. Initial testing showed a 95% reduction in manual validation time and caught 12 schema mismatches that would have caused downstream errors.