/simplify — Code & Document Optimization Review

Your preprocessing script works, but it's grown to 280 lines across four months and you know it has redundant logic.

When you need this

• After a working implementation is complete and you want to clean it up before sharing or publishing
• When a function has been patched multiple times and you suspect there's duplication
• After a document has gone through several revision rounds and feels bloated

What it does — and what it won't

/simplify reviews your code or document through five lenses: redundancy, complexity, performance, robustness, and readability. It identifies specific issues with line references and proposes concrete fixes. It applies changes only after you confirm.

Note: /simplify does not run tests. It reads the code and identifies issues — you should run your tests separately after applying any changes. Use it on working code, not broken code. After applying any fix, run your existing tests before committing — /simplify reads and proposes, but you verify.

Worked example

Priya's RNA-seq preprocessing script works correctly but has grown to 280 lines. She suspects there's duplicated logic across the filtering and normalization steps.

/simplify scripts/preprocess_rnaseq.R

/simplify returns five findings: the cell-filtering threshold check is written identically in three places (lines 44, 112, 201) — proposes a filter_cells() helper; a nested if/else block on lines 155–189 can be flattened to four guard clauses; the normalization loop re-reads the full count matrix on each iteration (O(n²)) — proposes a vectorized equivalent; a helper function defined but only called once; no check that the input matrix is non-empty before normalization. Priya reviews the report and decides which fixes to apply.

Try it

/simplify script.py
/simplify document.md