How to use the Python Anti-Patterns Checklist agent skill

What is an agent skill, and what does this one do?

An agent skill is a folder of instructions that an AI agent can load when a task matches the skill's description. This one is called python-anti-patterns and comes from the wshobson/agents repository, under plugins/python-development/skills/python-anti-patterns. It is released under the MIT licence.

Once the skill is active, the agent uses its file as a reference checklist when you ask it to review Python code. The checklist covers infrastructure, architecture, error handling, resource handling, type safety and testing. Each item shows a bad example, a suggested fix and a better example.

• Scattered timeout or retry logic that should be centralised
• Double retry, where both the application and the infrastructure retry
• Hard-coded configuration and secrets
• Exposed internal types such as ORM models leaking through APIs
• Mixed I/O and business logic
• Bare exception handling and swallowed errors
• Ignored partial failures in batch processing
• Unclosed files and other resources
• Blocking calls inside async code
• Missing type hints and untyped collections
• Tests that only cover the happy path and over-mocking

Install the skill from the ZIP download

Download the ZIP for this skill from this page. The skill is a single file, SKILL.md, so the archive is small. Unzip it so that the skill folder, containing SKILL.md, sits inside a skills directory. You can use either of these locations:

• Personal skills: ~/.claude/skills/ (available across your projects)
• Project skills: .claude/skills/ inside a specific project (shared with anyone who works in that repository)
• Claude apps: upload the skill under Settings → Capabilities → Skills

Trigger the skill with natural requests

You do not need a special command. The skill's description tells the agent when to use it: reviewing Python code for anti-patterns, using it as a checklist when reviewing code, before finalising implementations, or when debugging issues that might come from known bad practices. Ask in ordinary language and mention Python code and what you want checked.

Good requests name the file or change and the goal. For example, ask the agent to review a module before a merge, or to look for problems in a service that is misbehaving.

• Review this Python module for anti-patterns before I merge it.
• Use the Python anti-patterns checklist to check my API handlers for leaked database models.
• Before I finalise this async code, check for blocking calls.
• This batch job stops on the first error. Check it against common Python anti-patterns.
• Check this test file for happy-path-only coverage and over-mocking.

What to expect from a review

The agent compares your code with each section of the checklist and reports where it matches a listed anti-pattern. Because the skill is a written reference, the agent judges the code by reading it. It is not a parser or a scanner, so you should treat its output as a careful manual review rather than proof that the code is clean.

The file also ends with a quick review checklist you can work through yourself before finalising code, and a summary table pairing each anti-pattern with its fix.

• Expect findings that quote the relevant pattern and suggest the listed fix.
• Expect the agent to point to the good example style, such as context managers for files or specific exception types.
• Ask a follow-up if you want the fix applied to the code, or if a finding is unclear.

What you need and what the skill does not do

The skill lists no requirements. It does not run scripts and does not use the network, according to its listing. The only file in the skill is SKILL.md. Your agent environment must be able to load skills, and the code you review must be Python.

• Requires: nothing listed beyond an agent that can load skills
• Runs scripts: no
• Uses network: no
• Not for projects in other languages, because the checklist is Python-specific
• Not a replacement for a full linter, type checker or security scanner, because it does not analyse your whole codebase automatically

Tips for getting better results

Give the agent the whole relevant context. A single function may hide a problem such as double retry, which only shows up when you also look at the client configuration. The skill also points to a companion skill, python-design-patterns, for positive patterns and architecture guidance, so use it when you want to know what to build instead of what to avoid.

Pair the review with your normal tools. Run your linter, type checker and tests as usual, because this checklist covers judgement-based mistakes that those tools may miss.

• Include the surrounding code, such as the client or decorator used, when checking retry logic.
• Use the skill before merges and releases, or when you are teaching or setting team standards.
• Use the Quick Review Checklist at the end of the file as a self-check list.
• Keep the skill's file in your team's repository if you want everyone to review against the same standard.