Simplify
Skills to orchestrate coding agents to do long horizon tasks with little to no steering
npx -y skills add ronaknnathani/relay --skill simplifyAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Reduce complexity in recently changed code — cut needless abstraction, dead code, and duplication; improve names and structure — WITHOUT changing behavior. Quality only; it does not hunt for bugs (use review for that). Use after implementing a chunk of work, before review, or whenever a diff feels more complicated than the problem it solves. Triggers on "clean this up" / "tidy this" / "refactor this" when code already exists; for an underspecified goal with no code yet, use `clarify`.
SKILL.md
5.5 KB, as published. Nobody here has run it
Simplify
Make a change simpler and clearer without changing what it does. The bar: readable, explicit code over
clever, compact code — would a senior engineer call this overcomplicated? Scope is the recently
changed code — the working diff (e.g. git diff against the branch point) unless the caller names
files. This is quality-only: it does not hunt for bugs (that's review).
Step 0 — the invariant
Preserve functionality. Never change what the code does, only how it does it. Every original
feature, output, and behavior stays intact. If a simplification would alter behavior, it is a review
or implementation concern, not a simplification — stop and surface it.
Scan — three buckets
Look at the changed code for:
- Structural — nesting depth ≥ 3, functions longer than ~50 lines, nested ternaries, boolean-flag parameters that select behavior, indirection that only forwards a call.
- Naming —
data/result/temp/foo, names that mislead, names that imply something broader or different than the thing actually is. - Redundancy — duplicated logic, dead/unreachable code, unused parameters/variables/fields, over-abstraction (a helper, wrapper, or interface used in exactly one place), defensive code for states an upstream invariant already makes impossible.
The deletion test
For each abstraction or layer of indirection, ask: would removing it concentrate complexity (then keep it — it's earning its keep) or merely relocate it (then remove it — it's not)? A single-use helper usually relocates; inline it. A wrapper that hides a genuinely complex subsystem behind a small interface usually concentrates; keep it.
Two-sided guardrail
Simplifying is a balance, not a race to fewer lines. Pair every "look for" with a "don't":
| Look for | Don't |
|---|---|
| reduce nesting and indirection | introduce a nested ternary or a dense one-liner |
| inline a single-use helper | inline a helper that names a genuinely complex step |
| remove a redundant abstraction | remove a helpful abstraction that organizes the code |
| consolidate duplicated logic | merge unrelated concerns into one function |
| clearer names, fewer obvious comments | make the code harder to debug or extend |
If the "simplified" version is longer or harder to follow, it isn't a simplification — revert it.
Process
- Establish a baseline. Run the tests covering the change — or, if there is no suite or it cannot run, the narrowest available behavior check (build, typecheck, or lint) — and confirm it is green before you start. Never fabricate a green result; if the baseline is already red for unrelated reasons, note that and rely on the narrower check.
- Understand before changing (Chesterton's Fence). Work out why the code is shaped the way it
is — a guard, retry, or special case may exist for a real reason. Remove a guard only when its
premise is verifiable in the code you can see; if confirming it needs reasoning about whether an
upstream invariant truly holds, leave it and flag for
review. - One simplification at a time → re-run that check → keep it only if it stays green; revert immediately if not. Tests are necessary but not sufficient: for a change on a path the suite does not exercise, reason explicitly about input/output equivalence — error paths, ordering, side effects, short-circuit evaluation — before keeping it.
- Tests are out of scope. Never edit a test to make it pass — a breaking test means behavior changed, so revert the simplification, not the test. Do not refactor, consolidate, or delete tests or fixtures even when they look redundant; that silently weakens coverage.
- Keep the simplify diff separate from feature/fix changes, so a reviewer reads pure refactor.
- Verify functionality preserved at the end: the baseline check is green and you have reasoned about any untested paths you touched.
Red flags
- The simplified code is longer, or harder to follow, than before.
- Editing, consolidating, or deleting a test or fixture (tests are out of scope).
- Calling behavior "unchanged" from a passing suite alone, on code the suite does not exercise.
- Removing an abstraction or guard without applying the deletion test / Chesterton's Fence.
- Touching code outside the working diff ("while I'm here…").
- Optimizing for fewer lines at the cost of clarity.
Verification checklist
- The baseline check (tests for the changed code, or build/typecheck/lint) is green before and after — never a fabricated green.
- For any change on a path the suite does not cover, I reasoned about input/output equivalence, not just "tests pass".
- No test or fixture was edited, consolidated, or deleted — tests stayed out of scope.
- Every removed abstraction/guard passed the deletion test (and Chesterton's Fence for guards).
- The change stayed within the working diff.
- The diff is genuinely clearer — not merely shorter.
Gives 0 of the 12 instructions most refactoring skills give
Counted across 521 of the 525 authors here whose files we hold, read 2026-08-06
- run tests after each changein 59 of 521, across 56 files
- write tests before refactoringin 27 of 521, across 24 files
- preserve external behaviorin 26 of 521, across 22 files
- remove dead codein 25 of 521, across 24 files
- make small incremental changesin 20 of 521, across 17 files
- break the implementation into tiny commitsin 18 of 521, across 5 files
- ask the user about alternative optionsin 17 of 521, across 4 files
- create a GitHub issue with the planin 17 of 521, across 4 files
- explore the repository to verify assertionsin 17 of 521, across 4 files
- interview the user about the refactorin 16 of 521, across 3 files
- check the codebase for test coveragein 16 of 521, across 3 files
- refactor one thing at a timein 16 of 521, across 12 files
Said here and by no other author read
- reason explicitly about untested code paths
- verify functionality is preserved at the end
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once.