selenium-package

name	selenium-package
description	Package the output files from a selenium-pipeline run into a named zip and offer to attach it to a Jira ticket or save it to a custom path. Run this after selenium-pipeline completes for a directory.

Selenium package skill

Collects every artifact produced by a selenium-pipeline run for a given directory, zips them into a single named archive, then asks where to put it: attach to a Jira ticket, save to a custom path, or both.

Input

A directory name or path, e.g. courses or spec/selenium/courses. If not supplied, ask for one.

Optional --run-dir <root> — the run-scoped artifact root to package from. Defaults to tmp. The selenium-pipeline always passes this (e.g. tmp/selenium-runs/courses_20260602-141530); when set, discover artifacts under <root>/selenium-audit/, <root>/selenium-trim/, <root>/selenium-behavior/, and <root>/selenium-coverage/ instead of the bare tmp/ paths in Step 2.

Zip naming convention

<dir>_oplid-ws12_selenium_<ticket>_<YYYY-MM-DD>.zip

Examples:

courses_oplid-ws12_selenium_QE-51_2026-05-21.zip
grades_oplid-ws12_selenium_none_2026-05-22.zip

<ticket> comes from step 1 (the user supplies it, or none if skipped). <YYYY-MM-DD> is today's date from date +%Y-%m-%d.

Workflow

Step 1 — Collect inputs upfront

Ask both questions in one AskUserQuestion call:

1. Jira ticket key — the sub-task this pipeline run is tracked under (e.g. QE-51). Type none to omit from filename and skip attachment.
2. Destination — where to send the zip:
   • jira — attach to the Jira ticket
   • path — save to a custom local path (will prompt for it)
   • both — attach to Jira AND save locally

If the user picks path or both, ask for the target directory path in a follow-up question.

Step 2 — Discover artifact files

Look for pipeline output under these paths (using the short dir_name):

tmp/selenium-audit/<dir_name>.csv
tmp/selenium-audit/<dir_name>.summary.md
tmp/selenium-trim/<dir_name>.preview.md
tmp/selenium-trim/<dir_name>.manual_review.csv
tmp/selenium-trim/<dir_name>.follow_up.md
tmp/selenium-behavior/<dir_name>.behaviors.csv
tmp/selenium-behavior/<dir_name>.quality.csv
tmp/selenium-behavior/<dir_name>.quality.summary.md
tmp/selenium-behavior/<dir_name>.gap_report.md
tmp/selenium-coverage/<dir_name>/report.md
tmp/selenium-coverage/<dir_name>/run_current.log
tmp/selenium-coverage/<dir_name>/run_base.log
tmp/selenium-coverage/<dir_name>/trimmed.resultset.json
tmp/selenium-coverage/<dir_name>/original.resultset.json

For each path, check test -f <path> before including it. Files that don't exist are silently skipped (not every stage may have run). Report how many files were found before zipping.

If zero files are found, tell the user and stop.

Step 3 — Build the zip

TODAY=$(date +%Y-%m-%d)
TICKET=<ticket_key_or_none>
ZIP_NAME="<dir_name>_oplid-ws12_selenium_${TICKET}_${TODAY}.zip"
ZIP_PATH="tmp/${ZIP_NAME}"

zip -j "$ZIP_PATH" <file1> <file2> ...

Use -j (junk paths) so the zip contains flat files without nested directory structure, making it easy to open and read.

Report the zip path and size (ls -lh "$ZIP_PATH").

Step 4 — Deliver the zip

If destination is jira or both

Upload as an attachment to the ticket using the Jira REST API.

Find credentials — check these environment variables in order:

1. JIRA_BASE_URL and JIRA_API_TOKEN and JIRA_EMAIL
2. Fallback: inspect ~/.config/jira/config.yml or ~/.jira.d/config.yml

If credentials are not found in the environment, print clear instructions:

Could not find Jira credentials. Set these env vars and re-run:
  export JIRA_BASE_URL=https://your-domain.atlassian.net
  export JIRA_EMAIL=[e-mail]
  export JIRA_API_TOKEN=your_api_token

and stop the Jira upload (continue with local save if both was chosen).

Upload command:

curl -s -o /tmp/jira_attach_response.json \
  -w "%{http_code}" \
  -u "${JIRA_EMAIL}:${JIRA_API_TOKEN}" \
  -X POST \
  -H "X-Atlassian-Token: no-check" \
  -F "file=@${ZIP_PATH}" \
  "${JIRA_BASE_URL}/rest/api/2/issue/${TICKET}/attachments"

Check the HTTP status code:

• 200 — success. Parse the response JSON to get the attachment URL and show it to the user.
• 401 — bad credentials. Surface the error.
• 404 — ticket not found. Surface the error.
• Other — show the raw response.

If destination is path or both

Copy the zip to the user-supplied directory:

cp "$ZIP_PATH" "<target_dir>/${ZIP_NAME}"

Confirm the final path with ls -lh.

Step 5 — Report

Print a summary:

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  SELENIUM PACKAGE — <dir_name>
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  Archive:  <zip_name>
  Size:     <size>
  Files:    N artifacts

  Destinations:
  ✓ Attached to <ticket>
    <attachment_url>
  ✓ Saved to <target_path>
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Omit destination lines that were skipped or failed.

Notes

Flat zip (-j): All files land at the root of the zip with their original filenames. Two files could collide if they share a name (e.g. report.md doesn't exist here, but watch for it if new artifact types are added). If a collision is detected, use -j with explicit rename flags or switch to preserving paths with -r.

Large resultset files: trimmed.resultset.json and original.resultset.json are typically 6–7 MB each. Jira's default attachment limit is 10 MB per file, 25 MB total per request. If the zip exceeds 20 MB, warn the user before uploading and offer to omit the resultset files:

⚠ Zip is <size> — close to Jira's attachment limit.
  Include resultset JSON files? (They add ~13 MB but are needed to
  reproduce the coverage diff locally.)  y/n

tmp/ is a named Docker volume: the zip is built from files that were already docker cp'd to the host in the coverage-compare step. Do not try to zip from inside the Docker container — work on the host.