Skip to content
Back to skills

mutagate

ASecurity

Use mutation testing for test quality, inspect surviving mutants after writing tests, and check mutation scores before finishing tests.

  • 2 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 6, 2026
ai-agentskotlinshelltesting

Works with

  • cli

Security analysis

A100/100

Pro scans all 20 files and shows the line behind each finding

Scanned October 6, 2026

npx -y skills add keyboardsamurai/mutagate --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of mutagate?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for mutagate
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/keyboardsamurai-mutagate/badge)](https://www.skillsdirectory.com/skills/keyboardsamurai-mutagate)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: mutagate
description: Use mutation testing for test quality, inspect surviving mutants after writing tests, and check mutation scores before finishing tests.
---

# mutagate

Paths below are relative to this skill directory. Run commands with the project as the working directory, or pass `--repo <project>`.

If hooks are not installed (no `mutagate ... active` line at session start), ask the user, then run `scripts/mutagate install --agent <harness>` for the harness you run in.

After writing or changing a test file:

1. With hooks installed (a `mutagate ... active` line at session start), edits already register targets and start mutation runs in the background; skip to step 2. Without hooks, run `scripts/mutagate run <test-file>` to schedule a scoped mutation run.
2. Read the result with `scripts/mutagate check --wait 120`. It reuses finished background results, waits for running ones, and runs anything still stale, so don't start another run with `run --sync`. If it names a longer `--wait`, run exactly that command. `unscored: no mutants to test` means the target has nothing to mutate; nothing more is needed for it.
3. Improve assertions and edge cases for the surviving mutants; never weaken or delete production code to kill them. Read [the survivor playbook](references/survivor-playbook.md) for mutation-specific examples.
4. Run `scripts/mutagate check --wait 120` again after changing the tests (without hooks, repeat step 1 first).

Before reporting the task complete, run `scripts/mutagate check`. A nonzero exit means quality is below threshold, results are pending, tests are red or do not compile, or infrastructure needs attention; report the actual state rather than claiming a passing score.

If target resolution is ambiguous, select a listed target with `scripts/mutagate run <test-file> --target <target>`. The choice lasts for this session.

For equivalent mutants, use `scripts/mutagate waive <id> --reason "<reason>"` to propose a waiver. Never edit the waiver file or pass `--apply`; a human owns approval.

Run `scripts/mutagate status` to list survivors with the ids `waive` takes. Use `status --json` only for full results and filtered Kotlin mutants, and `scripts/mutagate doctor` to diagnose setup problems.

With hooks installed, the CLI joins the hook session automatically. Without hooks, these commands are advisory, and `MUTAGATE_SESSION` shares one session across shells. `<command> --help` prints its usage.

Files in this skill

  • AGENTS.md11 KB
  • CHANGELOG.md2.6 KB
  • CLAUDE.md11 B
  • CODE_OF_CONDUCT.md599 B
  • CONTEXT.md6.3 KB
  • CONTRIBUTING.md3.6 KB
  • SECURITY.md4.2 KB
  • SKILL.md2.4 KB
  • THIRD_PARTY.md1.9 KB
  • docs/eval_suite.md20.2 KB
  • docs/eval_suite_spec.md35.9 KB
  • docs/go_csharp_runners.md47.6 KB
  • docs/prd.md64 KB
  • docs/releasing.md8.8 KB
  • docs/validation.md10.5 KB
  • package.json1.2 KB
  • references/jars.json2 KB
  • references/kotlin-junk.yaml1.5 KB
  • references/result-schema.json2.7 KB
  • references/survivor-playbook.md2.4 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…