/test-review - Improve Test Assertions

SkillDev tools

Review tests for strict assertions and complete omitted validation. Use when the user wants to improve test assertions or mentions /test-review.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the /test-review - Improve Test Assertions skill

What this skill tells your AI

The instructions your AI receives, as published by rakovi4/continue-framework in .claude/skills/test-review/SKILL.md and read by ahel’s review.

Usage

  • /test-review — review all tests
  • /test-review {TestTarget} — review one test target

Available Templates

TemplatePurpose
.claude/templates/testing/test-review-patterns.mdUniversal rules, anti-pattern catalog, assertion rules
.claude/tech/{backend}/templates/testing/test-review-usecase.mdUsecase test patterns (Statements purity, 3-tier DSL)
.claude/tech/{backend}/templates/testing/test-review-rest.mdREST controller test patterns (mock matching)
.claude/tech/{backend}/templates/testing/test-review-storage.mdPersistence test patterns
.claude/tech/{backend}/templates/testing/test-review-acceptance.mdAcceptance test patterns (HTTP response assertions)
.claude/tech/{backend}/templates/testing/test-review-other.mdSelenium, email, scheduling, security patterns

Workflow

Scatter–gather: parallel read-only detectors find violations, then a single serial fixer applies them. The detectors never edit — only the fixer writes, one file at a time — so there are no concurrent writes to shared Statements files.

  1. Determine the test's layer (usecase / rest / storage / acceptance / selenium) — this selects the tech file each detector loads and whether the selenium detector runs.
  2. Dispatch the detectors concurrently and await all of them — start every named agent before awaiting results, then gather every result. No dispatch is detached or fire-and-forget: step 3 requires the complete result set. Each runs only its cluster of .claude/templates/workflow/test-review-checklist.md and returns a findings table:
    • test-review-assertions-agent — cluster A (assertion strictness)
    • test-review-placement-agent — cluster P (test-class vs Statements placement)
    • test-review-statements-agent — cluster S (Statements internal quality)
    • test-review-selenium-agent — cluster Se — only for selenium/frontend tests; skip for pure backend tests.
  3. Gather + fix: hand all findings to .claude/agents/test-review-agent.md (the serial fixer). It merges, dedups, applies fixes one file at a time, and prints the filled test-review-output-format.md. It runs /test-runner unless Stage 2 production work is concurrent; in that case the coordinator runs the joined test and implementation.

Small-file shortcut

If the target test is small with very few assertions, skip the fan-out and run a single test-review-agent pass over the whole checklist — the detector orchestration + merge overhead can exceed the single-agent cost on tiny files. Await that one dispatch too. This pass both finds and fixes, so an unawaited call could report the review done before a single assertion has been tightened.

Signals

GitHub stars
51
Forks
14
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
test-review-rakovi4
Source
github.com/rakovi4/continue-framework