Sparq:publish Results

Publish Test Results

Config, version check, and pattern rules per claude/rules/skills.md preamble.

Step 1: Resolve TMS Provider

Read outputs.tms.provider from sparq.config.json.

• If set: use that provider; confirm with user
• If missing: ask "Which TMS are you publishing to? (testrail / qase / zephyr)"

Step 2: Locate Test Output

Check common output paths in order:

1. playwright-report/results.json (Playwright JSON reporter)
2. test-results/*.json (Playwright blob reporter output)
3. test-results/**/*.xml (JUnit XML — Playwright or Cypress)
4. cypress/results/**/*.xml (Cypress JUnit XML)

If none found: ask the user to provide the path. Example prompt:

"I couldn't find test output at the default paths. Where is your test results file? (e.g., my-results/output.xml)"

Step 3: Detect Format

• File extension .json + contains "suites" or "stats" key → Playwright JSON
• File extension .xml + root element <testsuites> or <testsuite> → JUnit XML

Step 4: Parse Results

Playwright JSON (results.json):

• Iterate suites[].specs[].tests[]
• Extract: title (full test title), status (passed/failed/skipped/timedOut), duration (ms), error.message (if failed)

JUnit XML:

• Parse <testcase> elements
• Extract: name attribute (test title), presence of <failure> child (failed), <skipped> child (skipped), time attribute (seconds → ms), <failure> text (error message)
• timedOut detected from failure message containing "Timeout"

Step 5: Extract TC IDs

For each test result, apply regex TC-[A-Z0-9]+-[A-Z0-9]+-\d+ to the test title.

• Match found → associate result with that TC ID
• No match → group under "Untracked Tests" (separate run or appended section)

Step 6: Create Test Run in TMS

Create a test run named: SparQ Results — {date} {branch} (branch from git rev-parse --abbrev-ref HEAD if available, else omit).

• Include only the case IDs found in the parsed results (matched TC IDs resolved to TMS case IDs via tmsId frontmatter in .sparq/test-cases/)
• If no tmsId mappings exist, include all cases in the project/suite (provider default)

Step 7: Post Results per Test Case

Map SparQ/Playwright statuses to TMS statuses per <tms_tools> section. Post in bulk where supported (TestRail add_results_for_cases), per-result otherwise (Qase, Zephyr).

Step 8: Report

Emit summary:

[sparq] Results published:
  Provider: TestRail (project: {projectId})
  Run: {runName} — {runUrl}
  Posted: {N} passed, {N} failed, {N} skipped
  Unmatched (no TC ID): {N} tests (see "Untracked Tests" run)

Status Mapping

• passed → PASS (TestRail: 1, Qase: passed, Zephyr: PASS)
• failed / timedOut → FAIL (TestRail: 5, Qase: failed, Zephyr: FAIL)
• skipped → NOT_EXECUTED (TestRail: 3 (Skipped), Qase: skipped, Zephyr: NOT_EXECUTED)

TestRail

• Create run: mcp__testrail__add_run(project_id, name, case_ids[])
• Batch results: mcp__testrail__add_results_for_cases(run_id, results[]) — preferred for performance
• Per-result fallback: mcp__testrail__add_result_for_case(run_id, case_id, status_id, comment)
• comment field: include duration (ms) and error message truncated to 500 chars

Qase

• Create run: mcp__qase__create_run(project_code, title, description?)
• Per result: mcp__qase__create_result(project_code, run_id, case_id, status, time_ms, comment?)
• No batch API — loop per test result

Zephyr Scale

• Create cycle: mcp__zephyr__create_test_cycle(project_key, name, description?)
• Per result: mcp__zephyr__add_test_result(project_key, test_cycle_key, test_case_key, status, comment?)

References

• claude/skills/sparq-shared/references/tms-abstraction.md
• claude/skills/sparq-shared/references/mcp-tool-inventory.md
• claude/skills/sparq-shared/references/degradation-strategy.md

Usage

/sparq:publish-results
→ reads sparq.config.json (outputs.tms.provider: "testrail")
→ detects playwright-report/results.json (Playwright JSON format)
→ parses 42 test results, extracts 38 TC IDs
→ creates TestRail run "SparQ Results — 2026-02-20 feature/login"
→ posts 35 passed, 2 failed, 1 skipped via mcp__testrail__add_results_for_cases
→ output: https://team.testrail.io/index.php?/runs/view/42
→ 4 untracked tests (no TC ID) listed in summary

/sparq:publish-results
→ outputs.tms.provider: "qase"
→ detects test-results/junit.xml (JUnit XML format)
→ creates Qase run, posts results per mcp__qase__create_result
→ output: https://app.qase.io/run/PROJ/1