sniffable-python
Structure Python so LLMs can understand it in 50 lines.
npx skills add majiayu000/claude-skill-registry --skill sniffable-python-simhacker-moollm-3 --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Guides the agent to create Python code where the initial portion (first ~50 lines) includes a docstring containing purpose, usage, and examples, followed by an Argparse-based CLI tree. The implementation details are placed after the CLI structure, so the LLM can understand the tool from the opening notes. It emphasizes a canonical structure with shebang, docstring, imports, constants, and CLI definitions before any executable logic.
How it works
The skill prescribes:
- Use a shebang and a module docstring that becomes the --help output and is visible to the LLM.
- Place imports next, clearly grouped and visible at a glance.
- Define configuration constants (e.g., valid directions, max inventory) to expose API constraints up front.
- Build the CLI structure (via argparse) with subparsers and commands (e.g., move, examine, status), including arguments, options, and help text.
- Implement the actual behavior below the CLI fold, ensuring the LLM only reads the CLI tree for sniffability.
- Keep the implementation minimal or hidden behind internal functions like _do_move, _do_examine, _show_status, which are not needed by the LLM for understanding.
When to use it
Use when you need a Python script that is immediately understandable by an LLM from its opening lines, with a clear entry point (CLI) and usage examples, while keeping implementation details separate from the sniffable portion.
What it can touch
The skill allows tools: read_file, write_file. It structures code so the LLM reads the initial segment; actual file content touched by the agent should follow the prescribed CLI-first pattern and then the implementation.
Caveats
License: MIT. The approach assumes the presence of a standard Python CLI (argparse) and does not advocate introducing new syntax; it focuses on making existing syntax more readable to LLMs. It emphasizes avoiding decorator-based CLI frameworks in favor of an entirely separate CLI tree in main().
# Sniffable Python > **"Structure code for the first 50 lines. That's all the LLM needs."** Don't invent new syntax — **structure existing syntax** for optimal LLM comprehension. *If your code smells good, the LLM can sniff it from the first whiff.* *Steve Jobs wanted pixels you could lick. We want code you can sniff.* --- ## The Problem LLMs can read Python. But: - Large files overwhelm context - Implementation details bury the API - Comments scattered or missing - No clear entry point for understanding **Solution:** Structure Python so the first ~50 lines contain everything needed to understand and use the tool. Make your code's bouquet apparent from the opening notes. --- ## The Pattern ### Canonical Structure ```python #!/usr/bin/env python3 """skill-name: Brief description of what the script does. This docstring becomes --help output AND is immediately visible to the LLM. It should contain: - Purpose (one sentence) - Usage examples - Key behaviors Usage: python script.py command [options] Examples: python script.py move north --quiet python script.py examine sword --verbose """ import argparse from pathlib import Path import yaml # Configuration DEFAULT_ROOM = "start" VALID_DI
- The Problem
- The Pattern
- Canonical Structure
- The Zones
- Why This Works
- Dual-Audience Design
- DRY Principle
- Why Syntax Compression Fails
- The Symbol Collision Problem
- Comprehension Fidelity > Token Count
- Comments as YAML Jazz
- Comments: Substance Over Decoration
- Argparse vs Click vs Typer
- The Argparse Advantage
What does the sniffable-python skill do?
Structure Python so LLMs can understand it in 50 lines.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill sniffable-python-simhacker-moollm-3 --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.