agentleFS
naomiaronaomiaro/waveform-playlist/.claude/skills/mutation-testing/SKILL.md

mutation-testing

Skill1.7k starsChanged yesterdayMITView on GitHub

Use when checking that new or changed tests can fail - breaking a guarded line by hand, running mutants in bulk with the probes/mutate scripts, or deciding what a surviving, refused or hanging mutant means.

34 other projects use a version of this.

Trending

Quiet this year.thin dataabout the whole repository

Dec 2025: 0 mentionsJan 2026: 0 mentionsFeb 2026: 0 mentionsMar 2026: 0 mentionsApr 2026: 0 mentionsMay 2026: 0 mentionsJun 2026: 0 mentionsJul 2026: 0 mentionsAug 2026: 0 mentionsSep 2026: 0 mentionsOct 2026: 0 mentionsNov 2026: 0 mentions
Dec 2025Jun 2026Nov 2026
Mentions a month of the naomiaro/waveform-playlist repository that our weekly crawl found: only 11 in all, so this is a hint, not a trend. There is no adoption data for this file yet.

Most recently taken up by

Who's talking about naomiaro/waveform-playlist

The biggest 5 of 11 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 11

Talked about on GitHub, Stack Overflow and Hacker News (11 times).

On GitHub · 8

3 more on GitHub

On Stack Overflow · 2

On Hacker News · 1

Used with

Projects with it also use domain-modeling, code-review and 1 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • GoogleSkill

    domain-modeling

    Build and sharpen a project's domain model. Use when the user wants to pin down domain terminology or a ubiquitous language, record an architectural decision, or when another skill needs to maintain the domain model.

    Also in 3 projects that use this

  • GoogleSkill

    code-review

    Adversarial systems, concurrency, memory-model, and architecture code review protocol. Use this skill whenever asked to review code, PRs, diffs, or branches in Filament.

    Also in 5 projects that use this

  • MicrosoftSkill

    grill-with-docs

    Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 31 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Every new regression test is seen to fail with the line it guards broken, before the first review.

    line 8

“Common” is measured: kept by at least 6 in 10 of the 31 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Mutation checks

Every new regression test is seen to fail with the line it guards broken, before the first review.

Running mutants

  • Mutation checks in bulk: a small script that copies the file, applies one perl -0pi -e 's/…/…/', refuses to continue when the file did not change (a mutation that silently did not apply looks like a test that caught nothing), runs the one test file, prints the failing test names and restores the file. 33 mutations were run this way on #756; two showed an assertion gap (a flag that was never cleared, and a state value not tied to isRecording). Check git status --short for leftovers afterwards.
  • Mutation runs: give each mutant only the test files that can see its source file (40 s per mutant with 16 files, 3 s with the right 2 to 8; probes/mutate687.py, BY_FILE). Stop a run with pkill -INT -f mutate… so the finally restores the file; a background command is killed after 10 minutes without that.
  • The mutation runner over the full browser suite takes about 35 s per mutant (54 mutants, 30 minutes, in the background). Nothing else may edit source or run e2e against a dev server meanwhile: the server hot-reloads the mutated file. A mutant reported as REFUSED means prettier re-wrapped the line; fix the string and rerun it.
  • A mutant that removes a loop guard never ends, and vitest cannot time out a synchronous loop. Give the mutation runner a per-mutant timeout that kills vitest and counts the mutant as killed (probes/mutate689.py), and check pgrep -f vitest afterwards: a hung worker stays at 100% CPU.
  • A documentation agent reads source: do not run a mutation script while it works. Start the mutation run when the agent has returned, and read source yourself with git show HEAD:<file> while mutants are being applied.

Reading the result

  • A REFUSED mutant is not a killed one, and a fixture can hide the property. A pattern that matches twice is refused: make it unique and run it again. On #847 "states share the track object" passed with the memo removed, because the clip sat at tick 0, which a ppqn conversion leaves unchanged; and a rename test never looked at the old id. Choose fixture values the code under test changes. A survivor on a line no test can reach was a dead line or a redundant guard each time: remove it.
  • A mutant must be able to work before its survival means anything, and a survivor is first a question about the code. On #849 a "per id, not per object" mutant stored its value in a microtask, after the pass that would have read it: it survived for that reason alone. Of nine other survivors two were redundant code (a dev freeze of states that _emit freezes anyway; a guard before a call that already guards) and were REMOVED; the rest were missing assertions. Ask "would anything observable differ?" before writing a test for a survivor.
  • A mutant must break the property through the path the test reads. Two survivors on #770 added a duplicate name to a file the entry point does not export; the guard is about entry points. And a survivor on a line no caller can reach (a guard behind a disabled button) means the line goes, not that a test is missing.
  • A surviving mutant is not "equivalent" until the path is named. On #782 the one survivor of 67 (restAt clearing the range end) was reported as equivalent; the review found a path where the line mattered. For each survivor write down every caller that reaches the line and what it would see without it, or report it as an untested line.
  • A mutant must break the property, not resemble the old code. On #685 "the root renders again" did not make the parts render, because React skips a child whose element is the same object. The real break was two lines together (re-create the elements and drop the memo). When a mutant survives, ask what the test's subject actually depends on.
  • An option whose only test asserts that it was NOT used is untested. On #853 useCreateArrangement's one test with loadSource asserted that nothing was fetched, so deleting ...options passed every test (found by the final reviewer, who ran the suite against a copy of the hook). For each option of a new function, one test in which the option changes the outcome, and the mutant "the option is never passed".

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.