browser-agent-qa-testing

Name: browser-agent-qa-testing
Author: PramodDutta

182stars17forks

How to install browser-agent-qa-testing?

pramoddutta/qaskills/browser-agent-qa-testing

$npx -y skills add PramodDutta/qaskills --skill browser-agent-qa-testing

Installs into the current project.

Summary

Teach agents to use AI browser agents for exploratory and smoke QA with step budgets, evidence-based assertions, guardrails, and Playwright conversion.

Files · 1

View on GitHub

SKILL.md

1	# Browser Agent QA Testing Skill
2
3	You are an AI QA engineer who uses browser agents for bounded exploratory and smoke testing, gathers evidence for every claim, and converts stable findings into maintainable Playwright tests.
4
5	## Core Principles
6
7	1. Bound the agent: Define scope, credentials, data rules, and a maximum step budget before the run starts.
8	2. Require evidence: A browser agent must cite visible UI state, URL, network result, screenshot, or DOM observation.
9	3. Do not trust memory: Validate each important state in the live browser.
10	4. Protect data: Use test accounts, safe environments, and non-destructive workflows.
11	5. Prefer repeatable smoke paths: Use agents for discovery, then freeze stable paths into code.
12	6. Stop on uncertainty: If the agent cannot verify a result, it should report uncertainty instead of guessing.
13	7. Log decisions: Record why a path was explored, skipped, or converted to automation.
14	8. Avoid infinite browsing: Step budgets and charters keep exploration useful.
15
16	## Setup
17
18	Create a small harness for browser-agent QA runs.
19
20	```bash
21	python -m venv .venv
22	. .venv/bin/activate
23	pip install browser-use playwright pydantic python-dotenv
24	playwright install chromium
25	npm install --save-dev @playwright/test
26	```
27
28	Store run configuration outside prompts.
29
30	```text
31	qa-agent/
32	charters/
33	checkout-smoke.md
34	account-settings.md
35	evidence/
36	screenshots/
37	notes/
38	scripts/
39	run_browser_agent.py
40	tests/
41	e2e/
42	frozen-smoke.spec.ts
43	```
44
45	## Charter Template
46
47	Every agent run needs a charter.
48
49	```markdown
50	# Charter: Checkout Smoke
51
52	Goal: Verify a signed-in user can add one item to the cart and reach the payment step.
53	Environment: Staging
54	Account: Synthetic buyer
55	Step budget: 35
56	Allowed actions: Browse catalog, add item, open cart, start checkout
57	Forbidden actions: Submit real payment, change account email, delete saved addresses
58	Evidence required: Final URL, visible checkout heading, screenshot, console errors
59	Stop condition: Payment form is visible or a blocking bug is found
60	```
61
62	## Python Agent Runner
63
64	Keep the browser-agent task explicit and bounded.
65
66	```python
67	# qa-agent/scripts/run_browser_agent.py
68	import asyncio
69	from browser_use import Agent
70	from dotenv import load_dotenv
71
72	load_dotenv()
73
74	TASK = """
75	You are testing the checkout smoke charter.
76	Use the staging site only.
77	Do not submit payment.
78	Stop after 35 browser actions.
79	For every assertion, mention the exact visible text, URL, or screenshot evidence.
80	If blocked, report the blocker and stop.
81	"""
82
83	async def main() -> None:
84	agent = Agent(task=TASK)
85	result = await agent.run(max_steps=35)
86	print(result)
87
88	if __name__ == "__main__":
89	asyncio.run(main())
90	```
91
92	## Freeze to Playwright
93
94	Convert a stable exploratory path into deterministic automation.
95
96	```typescript
97	// tests/e2e/checkout-smoke.spec.ts
98	import { expect, test } from '@playwright/test';
99
100	test('signed-in buyer can reach payment step', async ({ page }) => {
101	await page.goto('/login');
102	await page.getByLabel('Email').fill(process.env.SMOKE_USER \|\| 'buyer@example.com');
103	await page.getByLabel('Password').fill(process.env.SMOKE_PASSWORD \|\| 'change-me');
104	await page.getByRole('button', { name: 'Sign in' }).click();
105
106	await page.getByRole('link', { name: 'Catalog' }).click();
107	await page.getByRole('button', { name: /add to cart/i }).first().click();
108	await page.getByRole('link', { name: 'Cart' }).click();
109	await page.getByRole('button', { name: 'Checkout' }).click();
110
111	await expect(page).toHaveURL(/checkout/);
112	await expect(page.getByRole('heading', { name: /payment/i })).toBeVisible();
113	});
114	```
115
116	## Evidence Rules
117
118	The browser agent report must include these fields.
119
120	1. Charter name.
121	2. Environment URL.
122	3. Account type.
123	4. Step count used.
124	5. Final URL.
125	6. Assertions with evidence.
126	7. Screenshots or trace links.
127	8. Console or network errors.
128	9. Bugs found.
129	10. Paths not covered.
130	11. Recommendation to automate or not automate.
131
132	## Guardrail Policy
133
134	Use guardrails to keep agent runs safe.
135
136	\| Guardrail \| Reason \| Example \|
137	\|---\|---\|---\|
138	\| Step budget \| Prevent wandering \| Stop at 35 actions \|
139	\| Test account \| Avoid customer data \| Synthetic buyer \|
140	\| Forbidden actions \| Prevent damage \| Do not submit payment \|
141	\| Evidence rule \| Reduce hallucination \| Cite visible text \|
142	\| Freeze criteria \| Create durable tests \| Convert stable smoke path \|
143	\| Human review \| Catch weak claim

People who liked this also viewed

6 picks

Type

tddSkills for Real Engineers. Straight from my .agents directory.SkillsJul 2026501k181k
test-driven-developmentUse when implementing any feature or bugfix, before writing implementation codeSkillsJul 2026175k259k
qaSkills for Real Engineers. Straight from my .agents directory.SkillsJul 2026160k181k
verification-before-completionUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidenceSkillsJul 2026153k259k
webapp-testingToolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser lSkillsJul 2026120k163k
playwright-best-practicesUse when writing Playwright tests, fixing flaky tests, debugging failures, implementing Page Object Model, configuring CI/CD, optimizing performance, mocking APIs, handling authentication or OAuth, teSkillsJul 202664k338

browser-agent-qa-testing

182stars17forks

View on GitHub

How to install browser-agent-qa-testing?

pramoddutta/qaskills/browser-agent-qa-testing

$npx -y skills add PramodDutta/qaskills --skill browser-agent-qa-testing

Installs into the current project.

Summary

Teach agents to use AI browser agents for exploratory and smoke QA with step budgets, evidence-based assertions, guardrails, and Playwright conversion.

Files · 1

View on GitHub

SKILL.md

1	# Browser Agent QA Testing Skill
2
3	You are an AI QA engineer who uses browser agents for bounded exploratory and smoke testing, gathers evidence for every claim, and converts stable findings into maintainable Playwright tests.
4
5	## Core Principles
6
7	1. Bound the agent: Define scope, credentials, data rules, and a maximum step budget before the run starts.
8	2. Require evidence: A browser agent must cite visible UI state, URL, network result, screenshot, or DOM observation.
9	3. Do not trust memory: Validate each important state in the live browser.
10	4. Protect data: Use test accounts, safe environments, and non-destructive workflows.
11	5. Prefer repeatable smoke paths: Use agents for discovery, then freeze stable paths into code.
12	6. Stop on uncertainty: If the agent cannot verify a result, it should report uncertainty instead of guessing.
13	7. Log decisions: Record why a path was explored, skipped, or converted to automation.
14	8. Avoid infinite browsing: Step budgets and charters keep exploration useful.
15
16	## Setup
17
18	Create a small harness for browser-agent QA runs.
19
20	```bash
21	python -m venv .venv
22	. .venv/bin/activate
23	pip install browser-use playwright pydantic python-dotenv
24	playwright install chromium
25	npm install --save-dev @playwright/test
26	```
27
28	Store run configuration outside prompts.
29
30	```text
31	qa-agent/
32	charters/
33	checkout-smoke.md
34	account-settings.md
35	evidence/
36	screenshots/
37	notes/
38	scripts/
39	run_browser_agent.py
40	tests/
41	e2e/
42	frozen-smoke.spec.ts
43	```
44
45	## Charter Template
46
47	Every agent run needs a charter.
48
49	```markdown
50	# Charter: Checkout Smoke
51
52	Goal: Verify a signed-in user can add one item to the cart and reach the payment step.
53	Environment: Staging
54	Account: Synthetic buyer
55	Step budget: 35
56	Allowed actions: Browse catalog, add item, open cart, start checkout
57	Forbidden actions: Submit real payment, change account email, delete saved addresses
58	Evidence required: Final URL, visible checkout heading, screenshot, console errors
59	Stop condition: Payment form is visible or a blocking bug is found
60	```
61
62	## Python Agent Runner
63
64	Keep the browser-agent task explicit and bounded.
65
66	```python
67	# qa-agent/scripts/run_browser_agent.py
68	import asyncio
69	from browser_use import Agent
70	from dotenv import load_dotenv
71
72	load_dotenv()
73
74	TASK = """
75	You are testing the checkout smoke charter.
76	Use the staging site only.
77	Do not submit payment.
78	Stop after 35 browser actions.
79	For every assertion, mention the exact visible text, URL, or screenshot evidence.
80	If blocked, report the blocker and stop.
81	"""
82
83	async def main() -> None:
84	agent = Agent(task=TASK)
85	result = await agent.run(max_steps=35)
86	print(result)
87
88	if __name__ == "__main__":
89	asyncio.run(main())
90	```
91
92	## Freeze to Playwright
93
94	Convert a stable exploratory path into deterministic automation.
95
96	```typescript
97	// tests/e2e/checkout-smoke.spec.ts
98	import { expect, test } from '@playwright/test';
99
100	test('signed-in buyer can reach payment step', async ({ page }) => {
101	await page.goto('/login');
102	await page.getByLabel('Email').fill(process.env.SMOKE_USER \|\| 'buyer@example.com');
103	await page.getByLabel('Password').fill(process.env.SMOKE_PASSWORD \|\| 'change-me');
104	await page.getByRole('button', { name: 'Sign in' }).click();
105
106	await page.getByRole('link', { name: 'Catalog' }).click();
107	await page.getByRole('button', { name: /add to cart/i }).first().click();
108	await page.getByRole('link', { name: 'Cart' }).click();
109	await page.getByRole('button', { name: 'Checkout' }).click();
110
111	await expect(page).toHaveURL(/checkout/);
112	await expect(page.getByRole('heading', { name: /payment/i })).toBeVisible();
113	});
114	```
115
116	## Evidence Rules
117
118	The browser agent report must include these fields.
119
120	1. Charter name.
121	2. Environment URL.
122	3. Account type.
123	4. Step count used.
124	5. Final URL.
125	6. Assertions with evidence.
126	7. Screenshots or trace links.
127	8. Console or network errors.
128	9. Bugs found.
129	10. Paths not covered.
130	11. Recommendation to automate or not automate.
131
132	## Guardrail Policy
133
134	Use guardrails to keep agent runs safe.
135
136	\| Guardrail \| Reason \| Example \|
137	\|---\|---\|---\|
138	\| Step budget \| Prevent wandering \| Stop at 35 actions \|
139	\| Test account \| Avoid customer data \| Synthetic buyer \|
140	\| Forbidden actions \| Prevent damage \| Do not submit payment \|
141	\| Evidence rule \| Reduce hallucination \| Cite visible text \|
142	\| Freeze criteria \| Create durable tests \| Convert stable smoke path \|
143	\| Human review \| Catch weak claim

People who liked this also viewed

6 picks

Type