Evidence Collector

Validates web implementations against specifications using Playwright screenshot evidence and interactive testing.

2|Updated May 21, 2026
One-click install
npx skills add https://github.com/tcvdog/agency-agents-hermes --skill evidence-collector-tcvdog
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: Evidence Collector
Source: https://github.com/tcvdog/agency-agents-hermes/tree/main/testing/evidence-collector
Command: npx skills add https://github.com/tcvdog/agency-agents-hermes --skill evidence-collector-tcvdog

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? QA reviews often accept unverified claims like "zero issues found" or "production ready" without proof. This Skill enforces evidence-based quality assurance by requiring screenshots and documented test results for every claim, catching broken functionality before approval. ## Core Features & Use Cases - Visual Evidence Capture: Runs Playwright screenshot scripts across desktop, tablet, and mobile viewports plus dark mode to document actual rendered output. - Interactive Element Testing: Verifies accordions, forms, navigation, mobile menus, and theme toggles actually function, with before/after screenshot comparisons. - Specification Compliance Checking: Quotes the original spec and compares it line-by-line against what screenshots actually show, flagging fantasy reporting and unrequested "luxury" claims. - Use Case: After a developer delivers a landing page claiming premium design and full responsiveness, run this Skill to capture screenshots at multiple resolutions, test all interactive elements, and produce a report listing the 3-5 real issues that must be fixed. ## Quick Start Use the evidence-collector skill to QA the site running at localhost:8000 and produce an evidence-based report with screenshots.

Frequently Asked Questions about Evidence Collector

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I QA a website with screenshot evidence?▼

Run the Playwright capture script against the local server URL to generate desktop, tablet, and mobile screenshots, then compare each capture against the original specification text. Document every mismatch with the specific screenshot filename as proof.

How to test accordions and forms with Playwright screenshots?▼

Capture before and after screenshots of each interactive element, such as accordion-0-before.png versus accordion-0-after.png. If the images are identical, the interaction failed; check test-results.json for TESTED or ERROR status per element.

Does this QA approach support dark mode and mobile testing?▼

Yes, the capture script generates dark-mode screenshots and responsive captures at 1920x1080, 768x1024, and 375x667 resolutions. It also verifies theme toggle switching between light, dark, and system modes.

Why does the QA report default to finding 3-5 issues?▼

First implementations realistically always contain defects, so a zero-issues result signals insufficient inspection rather than perfect code. The methodology treats perfect scores and unverified luxury claims as automatic failure triggers requiring deeper review.

What are the limitations of screenshot-based QA testing?▼

Screenshot comparison verifies visual and interactive behavior but cannot assess backend logic, security, or performance under load. It also requires a running local server and the Playwright capture script to be present in the project.