All Agents Overview

swayambhu-qa ships 8 agents. Each handles a specific job in the QA lifecycle. Use the full pipeline for end-to-end runs, or call individual agents when you need just one step.

When to use which agent

Agent	Starts from	Use when
/qa-pipeline	Requirement ticket	Full run — ticket → TCs → automation → results → PR
/create-test-cases	Requirement ticket	QA lead wants to write & review TCs before automating
/generate-tests	A live URL	Quick automation from any URL — no ticket or TCs needed
/automate-from-tms	Existing TCs in TMS	TCs are already in Xray/TestRail — just need automation code
/heal-tests	Failing spec file	Tests broke after a UI change — fix selectors automatically
/bug-to-test	Bug report	Turn a confirmed bug into a regression test
/analyze-flaky	Test results JSON	Identify which tests are flaky and why
/qa-report	Test results JSON	Generate a shareable report to post or present

Agent details

/qa-pipeline

The full orchestrator. Runs all phases from ticket fetch to Draft PR.

/qa-pipeline --id TEST-22 --source jira --tool playwright --tms xray
/qa-pipeline --id QA-42 --source jira --tool playwright,restassured
/qa-pipeline --id 42 --repo myorg/myrepo --no-pr

→ Full reference

---

/create-test-cases

Reads a ticket and generates test case documents only — no automation code written. Good for teams that review TCs in Xray before writing any automation.

/create-test-cases --id TEST-22 --source jira --tms xray

After TCs are approved in Xray, run /automate-from-tms to generate the code.

→ Full reference

---

/generate-tests

Takes a live URL, scrapes the page, and writes a Playwright spec file directly — no ticket or TCs needed.

/generate-tests https://myapp.com/login
/generate-tests https://myapp.com/checkout --screenshot

→ Full reference

---

/automate-from-tms

Reads existing test cases from your TMS (Xray, TestRail, or local markdown files) and generates automation code, runs it, heals failures, and logs bugs.

/automate-from-tms --id QA-42 --source jira --test-mgmt xray --tool playwright
/automate-from-tms --suite "Login Tests" --test-mgmt testRail --tool playwright,restassured
/automate-from-tms --case TC-1-01,TC-1-03 --test-mgmt markdown --tool playwright

→ Full reference

---

/heal-tests

Auto-fixes broken selectors and timing issues in a failing spec file. Skips failures that are real app bugs (those get escalated to bug reports).

/heal-tests tests/generated/login.spec.ts

→ Full reference

---

/bug-to-test

Takes a bug report (issue ID or description) and writes a regression test that fails on the bug and passes after it's fixed.

/bug-to-test --id BUG-123 --source jira

→ Full reference

---

/analyze-flaky

Reads test results across multiple runs and identifies flaky tests, their failure rate, and the likely cause.

/analyze-flaky reports/results-TEST22.json

→ Full reference

---

/qa-report

Generates a structured QA report from test results — suitable for posting to JIRA, Slack, or a PR comment.

/qa-report --id TEST-22 --source jira

→ Full reference

---

Multi-run mode

Four agents support running against multiple issues or files in a single command. The others don't — here's why.

Which agents support multi-run

Agent	Multi-run?	Example
/qa-pipeline	✅ Yes	--id TEST-22,TEST-62
/create-test-cases	✅ Yes	QA-42,QA-43,./local-spec.txt
/automate-from-tms	✅ Yes	--id QA-42,QA-43,./test-cases/TC-login.md
/bug-to-test	✅ Yes	--jira BUG-101,BUG-102 --file ./bug3.txt
/heal-tests	❌ No	—
/analyze-flaky	❌ No	—
/generate-tests	❌ No	—
/qa-report	❌ No	—

Why only these four?

/qa-pipeline, /create-test-cases, /automate-from-tms, /bug-to-test all share the same input model: they take an issue ID or file path as their primary input. Repeating that input is a natural and common need — a QA lead wants to automate three tickets at once, or a team has a batch of bug reports to convert to regression tests. Multi-run is simply "run the same agent N times, sequentially, with results collected."

The other agents have a fundamentally different input model:

• /heal-tests — takes a failing spec file, not an issue ID. It already handles every failing test within that spec in one run. Running it across multiple spec files is handled by the agent naturally, not by a separate multi-run loop.

• /analyze-flaky — takes a test results JSON. Flakiness analysis is cross-run by design: the agent reads multiple execution histories from one results file and reasons across all of them. There is no meaningful "batch" version — you aggregate results first, then analyze.

• /generate-tests — takes a live URL. Multi-URL support is possible but rarely needed: you'd normally produce one focused spec per URL. For multiple pages, call the agent once per URL.

• /qa-report — aggregates test results into one report. Reports are already a many-to-one operation — one report covers all tests in a run. Running the reporter twice against different inputs would produce two separate reports, not a combined one.

How multi-run works

All four agents run sequentially with a 5-second cooldown between items:

🗂️  Multi-run — 3 items queued: QA-42, QA-43, ./local-spec.txt
    Processing sequentially (5s cooldown between items).

--- Item 1 of 3: QA-42 ---
[... full agent run ...]
✅ QA-42 complete. Starting next in 5 seconds...

--- Item 2 of 3: QA-43 ---
[... full agent run ...]
✅ QA-43 complete. Starting next in 5 seconds...

If one item fails (ticket not found, file missing), it is marked ❌ and the agent continues with the next. A combined summary is printed after all items complete.

Mixed mode — ticket IDs + file paths in one command

All four agents accept a mix of issue IDs and local file paths in the same command. File paths are detected automatically by prefix (./, /) or extension (.md, .txt, .docx, .doc, .pdf). No IMS credentials are required for file entries.

# Two JIRA tickets + one local spec file
/qa-pipeline --id "TEST-22,TEST-62,./docs/my-feature.md" --source jira --tool playwright

# JIRA ticket + local file (no credentials needed for the file entry)
/create-test-cases "QA-42,./local-spec.txt" --source jira

# Ticket + local TC markdown file
/automate-from-tms --id "QA-42,./test-cases/TC-login.md" --source jira --test-mgmt xray --tool playwright

# JIRA bug + local bug report file
/bug-to-test --jira BUG-101 --file "./bugs/manual-bug.txt"

---

All command combinations

Every agent, every --tool, --source, and --tms option in one place → Command Combinations