AgentSkillsCN

recurse-ml

运用 Python 实现机器学习编码模式，有效识别并规避常见陷阱。触发条件：在遵循 Python 最佳实践编写机器学习代码时。

SKILL.md
--- frontmatter
name: recurse-ml
description: >
  Python ML coding patterns for identifying and avoiding common pitfalls.
  Trigger: When writing ML code with Python best practices.
license: Apache-2.0
metadata:
  author: poletron
  version: "1.0"
  scope: [root]
  auto_invoke: "Working with recurse ml"

## When to Use

Use this skill when:
- Writing Python ML/data science code
- Identifying common Python anti-patterns
- Improving code reliability and readability
- Avoiding subtle bugs in ML pipelines

Decision Tree

code
Need exception handling?   → Use specific exceptions (bare-exceptions.md)
Need conditionals?         → Check conditionals.md for patterns
Need boolean checks?       → See bool.md for comparisons
Need type safety?          → Apply typing.md guidelines
Need debugging?            → Use rml-verify.md

Critical Patterns

Don't Catch Bare Exceptions (REQUIRED)

python
# ❌ BAD - Hides unintended exceptions
try:
    risky_operation()
except:
    handle_error()

# ❌ EQUALLY BAD
try:
    risky_operation()
except Exception:
    handle_error()

# ✅ GOOD - Catch specific exceptions
try:
    risky_operation()
except SpecificException:
    handle_error()

# ✅ OK if reraising
try:
    risky_operation()
except SpecificException as e:
    handle_error(e)
    raise  # Reraise the exception

Why: Bare exceptions hide bugs and give false stability.

Resources

Specialized ML coding patterns in this skill: