agentleFS
desbmadesbma/agent-skills/review-tests/SKILL.md

review-tests

Skill1 starsChanged 16 days agoMITView on GitHub

Review Jujutsu revision changes from jj show and propose test additions, fixes or removals.

9 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 5 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 5

    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:

    Follow the common review instructions: @../review-common.md

    line 12
  2. Distinctive:

    test behavior that is left untested and would regress silently

    line 18
  3. Distinctive:

    remove a test that is useless: it cannot fail, it duplicates another one, or it exercises the language and its libraries rather than the …

    line 21
  4. Distinctive:

    Never propose to change the main code to make it easier to test, take it as it is

    line 27
  5. Distinctive:

    Prefer a few meaningful tests over coverage for its own sake

    line 28

“Common” is measured: kept by at least 6 in 10 of the 8 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

Review tests of Jujutsu changes

Review the Jujutsu changes for the target revision, and propose changes to the tests covering them.

Follow the common review instructions: @../review-common.md

Main goals

Propose to:

  • test behavior that is left untested and would regress silently
  • pin down a corner case whose handling looks intentional but not obvious, so that the test documents the intent
  • fix a test that is incomplete, or asserts the current behavior rather than the intended one
  • remove a test that is useless: it cannot fail, it duplicates another one, or it exercises the language and its libraries rather than the code
  • flag a test that manufactures a state the production code cannot reach: it pins a special case that exists only to satisfy the test, and the pair should go together
  • flag a test that can turn flaky: hardcoded sleep or timing values, assumptions about the host system (paths, locale, clock, available ports), or unsynchronized concurrency racing on shared state

Guidelines

  • Never propose to change the main code to make it easier to test, take it as it is
  • Prefer a few meaningful tests over coverage for its own sake
  • State the regression each proposed test would catch
  • Stay on tests: correctness, readability and documentation are covered by other review domains

Severity

A branch is an alternative the code can take: an arm of an if, a match or a conditional expression, an exception handler, a loop body, the right operand of a short-circuit and or or, a function's body. A test checks a branch when it fails once that branch is deleted.

Rate the defect of the suite the item reports by the signal the suite gives:

  • critical — a false signal: an existing assertion expects a result the code should not produce, or a test fails on a state the code handles
  • major — no signal: a branch the revision adds is checked by no test
  • minor — any other defect: a branch the revision did not add is checked by no test, a value a checked branch computes goes unasserted, a test is redundant or useless

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.