This repository offers a collection of reusable Claude Code slash command skills designed to automate various software development tasks. Targeting developers utilizing Claude Code, these skills aim to enhance code quality, improve documentation, and streamline testing processes by leveraging LLM capabilities for code analysis and generation.

How It Works

The project comprises skills implemented as markdown files (.md.txt). Users install a skill by copying its content to ~/.claude/commands/ and renaming it to .md. By default, skills are configured with disable-model-invocation: true to prevent accidental LLM context pollution. A key security feature is the .md.txt file extension, which ensures that potentially malicious HTML comments, hidden from markdown previews, are visible in the raw file content before installation. Claude Code operates with broad system access, making skill review critical.

Quick Start & Requirements

• Install: Copy skill file to ~/.claude/commands/ and rename from .md.txt to .md.
• Prerequisites: Claude Code environment.

Highlighted Details

• tighten-types: Systematically reviews Python source files to improve type annotations.
• contract-docstrings: Generates docstrings detailing function contracts (preconditions, raises, silences).
• hypothesis-tests: Creates property-based tests using Hypothesis by analyzing production code.
• mutation-testing: Assesses test suite strength by introducing deliberate bugs and checking test coverage.
• pre-mortem: Identifies fragile code areas and generates hypothetical future bug post-mortems.
• stub-package: Generates a condensed structural overview of Python packages/modules.
• try-except: Audits try/except blocks for overly broad scope and potential issues.

Licensing & Compatibility

• License: MIT.
• Compatibility: No explicit restrictions noted for commercial use.

Limitations & Caveats

Skills represent "maximum privilege remote code execution" due to Claude Code's broad filesystem, shell, and credential access. Users must meticulously review the raw .md.txt file contents for hidden malicious instructions within HTML comments before installation, as rendered markdown previews can obscure such content.