agentleFS
reliable-coding-agentsreliable-coding-agents/reliable-python/skills/limiting-nesting/SKILL.md

limiting-nesting

Skill1 starsChanged 2 months agoMITView on GitHub

Use when writing or reviewing any Python function that nests blocks — if inside for inside try, guard pyramids, arrow-shaped code. Keeps block depth at four levels or fewer by extracting units and returning early.

The file

Files
SKILL.md

Limiting Nesting

Keep a function at four levels of block nesting or fewer. The function body is level one. Past level four, extract a named unit or invert a condition.

Depth is not a style preference. Each level is a condition the reader must hold in their head to know whether a line runs at all. At five levels the reader is tracking five simultaneous truths, and the line that finally does the work is 40 characters from the left margin.

When to Use

Load this before writing or editing:

  • Any if inside a for inside a try
  • Validation that grows a new if per field
  • A loop body that filters, then branches, then acts
  • Any block whose closing lines are a run of dedents
  • Code where the real work sits at the far right of the screen

When NOT to apply: depth is a signal, not a defect. A parser or state machine with a genuinely four-deep decision structure is fine at four. The rule draws the line where a reader stops being able to enumerate the conditions, not where the code stops being pretty.

Core Pattern

# Nests 5 levels: for -> if -> for -> if -> the actual work
def collect_errors(reports):
    findings = []
    for report in reports:
        if report.enabled:
            for entry in report.entries:
                if entry.severity == "error":
                    findings.append(entry)
    return findings

Two independent fixes. Invert to return early, which removes a level at the top:

def collect_errors(reports):
    findings = []
    for report in reports:
        if not report.enabled:
            continue
        for entry in report.entries:
            if entry.severity == "error":
                findings.append(entry)
    return findings

Or extract the inner unit, which gives the inner logic a name:

def _error_entries(report):
    return [entry for entry in report.entries if entry.severity == "error"]


def collect_errors(reports):
    findings = []
    for report in reports:
        if report.enabled:
            findings.extend(_error_entries(report))
    return findings

The second is usually better: _error_entries is independently testable, and the outer function now reads as one sentence.

Techniques

ShapeFix
if cond: wrapping the whole bodyInvert it — if not cond: return
if inside a loop wrapping the bodyif not cond: continue
Nested loop doing its own filteringExtract the inner loop as a named function
try wrapping a large blockWrap only the call that raises
if/elif ladderAlready flat — an elif chain counts as one level
Filter, then branch, then actComprehension for the filter, loop for the act
Deep if on optional fieldsValidate and return early at the top of the function

An elif ladder is deliberately not a violation. if/elif/elif/else reads as one flat decision, and the checker counts it as one level even though the AST nests it.

Common Mistakes

Extracting a fragment instead of a unit. Pulling out three lines that need five parameters and only make sense at that one call site trades nesting for coupling. If you cannot name the extracted function without saying "part two", invert a condition instead.

Flattening by combining conditions. if a and b and c and d: is one level and unreadable. The nesting was a symptom; a four-clause condition is the same complexity wearing a different shape.

Hiding depth in a comprehension. A triple-nested comprehension is not flatter than the loop it replaced, it is the same depth with fewer places to put a name.

Rationalizations

ExcuseReality
"It's only one level over"Then one continue fixes it. The cheapness of the fix is the argument for making it.
"Extracting would hurt performance"A Python function call is nanoseconds. Measure before trading readability for it.
"The logic genuinely is nested"Genuinely nested logic is exactly what deserves named intermediate steps.
"It's a short function"Depth and length are different problems. A 12-line function can be six levels deep.
"I'd need to pass too many arguments"That is the finding, not an objection to it. High argument count means the block was reading state it should not have been.

Red Flags — Stop and Flatten

  • The closing lines of a function are a staircase of dedents
  • You scroll right to read the working line
  • A loop body is entirely one if
  • You are about to add a level to code that already has four
  • A block's purpose needs a comment because its conditions are off-screen

Checker Rule

Reported as CS01 (Long Method), whose catalog signal explicitly includes deep nesting. POT04 covers the other half — a function over 60 code lines — so the two together cover both ways a function becomes hard to hold in the head.

CS01: function collect_errors nests 5 levels deep (limit: 4)
  remedy: Extract the inner block into a named function, or invert a
          condition to return early and remove a level.

The finding anchors to the deepest statement, not the function definition, so it points at the line that needs to move.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.