agentleFS
zaxbysaucezaxbysauce/opencode-swarm/.agents/skills/test-file-split/SKILL.md

test-file-split

Skill517 starsChanged 3 days agoMITView on GitHub

Protocol for splitting test files that approach or exceed the FR-006 500-line limit (enforced in CI by scripts/check-test-file-cap.ts as a diff-scoped ratchet). Covers describe-block extraction, shared helper management, pure-function extraction, mock isolation verification, and cascading-split detection. Load when a test file approaches or exceeds 500 lines.

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 2 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read first: .opencode/skills/writing-tests/SKILL.md (or .claude/skills/writing-tests/SKILL.md) for bun:test framework rules, mock isolation …

    line 12
  2. Very specific:

    Create tests/helpers/<module>-shared.ts with shared fixtures, mock factories, and setup functions.

    line 46
  3. Very specific:

    If only bun:test imports and 1-2 source imports are shared, duplicate them in both files.

    line 49
  4. Very specific:

    Before renaming or moving shared code as part of the split, size the blast radius with the repo_map tool: run repo_map with action …

    line 53
  5. Distinctive:

    Add all necessary imports to the new file.

    line 63
  6. Distinctive:

    Remove now-unused imports from the original file.

    line 64
  7. Very specific:

    If the co-run fails but isolated runs pass, check for mock.module() leakage.

    line 87
  8. Very specific:

    After splitting, evaluate whether internal utility functions in the source module can be exported via _test_exports for zero-mock testing.

    line 91

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “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

Test File Split Protocol (FR-006)

scripts/check-test-file-cap.ts enforces the 500-line cap per test file (FR-006 / SC-006.1) as a diff-scoped ratchet: new test files over 500 lines and existing over-cap files that grew fail the quality gate and block PR merge. Pre-existing over-cap files not touched by the PR are non-blocking. Escape hatch: TEST_CAP_ENFORCE=0 soft-warns. This skill covers the complete splitting protocol.

Read first: .opencode/skills/writing-tests/SKILL.md (or .claude/skills/writing-tests/SKILL.md) for bun:test framework rules, mock isolation patterns, and file placement conventions.

When to use this skill

  • A test file exceeds or approaches 500 lines
  • CI fails with an FR-006 file-size violation
  • You are adding tests to a file that is already above 400 lines (proactive split)

Step 1 — Measure and identify split boundaries

# Check the file
wc -l tests/unit/scripts/my-module.test.ts

# Find all test files exceeding 400 lines (early warning)
find tests/ -name "*.test.ts" -exec wc -l {} \; | sort -rn | awk '$1 > 400'

Identify natural describe() block boundaries. Group blocks by functional area:

  • Each describe() block should belong to exactly one split file
  • Shared beforeEach/afterEach hooks determine which blocks must stay together

Step 2 — Choose a suffix for the new file

PatternExampleWhen to use
<module>-<area>.test.tsrelease-notes-fragments-sha.test.tsSplit by functional area (SHA resolution, validation, merge logic)
<module>-<area>.adversarial.test.tsauth-login.adversarial.test.tsSplit adversarial tests into their own file

Step 3 — Manage shared imports and helpers

Three options, in order of preference:

  1. Extract to shared utility (preferred for complex shared setup): Create tests/helpers/<module>-shared.ts with shared fixtures, mock factories, and setup functions. Import from both split files.
  2. Duplicate simple imports (for small overlap): If only bun:test imports and 1-2 source imports are shared, duplicate them in both files. Simpler than a utility module for trivial cases.
  3. Extract pure functions from source (for testability): If the source module has inline validation logic, extract them as exported pure functions (e.g., isValidPrNumber, resolveAllCandidates) so both test files can target them independently. See the PR #1762 example below. The extraction pattern: move the inline predicate or validator into an exported pure function and import it from both test files. Before renaming or moving shared code as part of the split, size the blast radius with the repo_map tool: run repo_map with action importers on the target file to find every consumer, repo_map (action: dependencies) to see what the target itself needs, and repo_map with action blast_radius to enumerate transitive dependents before you move anything.

Step 4 — Extract and move describe blocks

  1. Cut the selected describe() blocks from the original file.
  2. Paste them into the new file.
  3. Add all necessary imports to the new file.
  4. Remove now-unused imports from the original file.

Step 5 — Verify both files

Line count check

wc -l tests/unit/scripts/my-module.test.ts tests/unit/scripts/my-module-sha.test.ts
# Both must be under 500 lines

Isolated run

bun --smol test tests/unit/scripts/my-module.test.ts --timeout 60000
bun --smol test tests/unit/scripts/my-module-sha.test.ts --timeout 60000

Co-run (mock isolation verification)

# Critical: Bun shares a single process across test files.
# mock.module leaks can cause co-run failures even when isolated runs pass.
bun --smol test tests/unit/scripts/my-module*.test.ts --timeout 60000

If the co-run fails but isolated runs pass, check for mock.module() leakage. See .opencode/skills/writing-tests/SKILL.md → "Mock Isolation Rules" and the _internals DI seam pattern.

Step 6 — Evaluate _test_exports opportunity

After splitting, evaluate whether internal utility functions in the source module can be exported via _test_exports for zero-mock testing. This is a natural cleanup moment — the split already forces you to review test coverage boundaries.

Cascading split warning

If a previously split file exceeds 500 lines again, the test suite is structurally too large for a single module. Do not split a third time — reorganize the tests by source module boundaries instead. Repeated splitting produces fragmented test suites that are hard to navigate and maintain.

Real-world example (PR #1762)

tests/unit/scripts/release-notes-fragments.test.ts exceeded 500 lines. It was split into:

FileLinesContent
release-notes-fragments.test.ts379Fragment collection, deduplication, output formatting
release-notes-fragments-sha.test.ts261extractCommitShasFromBody, mergeCandidateLists, resolveAllCandidates, isValidPrNumber, stripCustomReleaseNotesBlock

The split also extracted isValidPrNumber and resolveAllCandidates as pure exported functions from the source module, enabling independent testing.

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.