TestivAI MCP — Watchlist

MCP100 Assessment

Watchlist / Tracking

Why we’re tracking it

Follow-up value is specific to TestivAI visual-regression result inspection: Do verdicts still distinguish identical DOM/style render noise from structural and style-only changes?

Reviewed Oct 5, 2026 · Documentation review; complete setup workflow not independently tested

Best fit

Local-first visual regression evidence and diff inspection for AI coding agents.

What we know

MCP server in the TestivAI OSS suite. It exposes visual test results and diff images so the same coding agent that changed a UI can inspect regressions locally. Baseline approval is intentionally a human decision.

Maintenance / activity evidence

@testivai/mcp 0.6.0 published 2026-10-05T00:07:04.513Z

Tracking evidence observed Oct 5, 2026. These dated signals do not establish current maintenance. Reviewed enrichment confidence: high.

Verification gaps / watch-outs

Installation and comprehensive security have not been independently audited.
Read local visual-test evidence; approval should not be automated by agent.
Authentication not yet verified.
Sensitive access: Screenshots/DOM snapshots may contain application data

Primarily read/verification; biggest safety rule is not allowing the agent to self-approve baselines.

Full authority, access and evidence limits are available in Review context and Access & risk.

All limitations / Review context

Do verdicts still distinguish identical DOM/style render noise from structural and style-only changes?

The bounded review of the publisher's documented TestivAI visual-regression result inspection and release notices identified no disclosed material condition preventing an accurately scoped public tracking entry. The resource's operational permissions remain uncertainty; this assessment does not certify installation safety.

Primarily read/verification; biggest safety rule is not allowing the agent to self-approve baselines.
Read local visual-test evidence; approval should not be automated by agent.
Visual reports/diffs stay local in the OSS workflow.
Sensitive access: Screenshots/DOM snapshots may contain application data
Installation, client compatibility and comprehensive security have not been independently audited. Authentication and access depend on the configured services.
Runtime behavior, dependencies and undisclosed vulnerabilities were not independently audited.
Documented tool authority and any remote data handling must be checked before installation or use.
Tool inventory, client matrix and permissions completeness were not requirements for this Radar tracking review.
Baseline approval must remain human-controlled.
Dynamic content can create false diffs unless ignored/tolerated.
Exact MCP tool inventory can evolve with the suite.
No hosted AI service is needed for local diffing.
Do not auto-approve visual baselines; project explicitly keeps approval human-controlled.

Original activity observations

@testivai/mcp 0.6.0 published 2026-10-05T00:07:04.513Z

Technical identity

npm/@testivai/mcp

Exact implementation source: https://www.npmjs.com/package/@testivai/mcp ↗
Reviewed resource type: Testing / Debugging
Repository identity: testivai/testivai-oss
Package identity: npm:@testivai/mcp

Evidence & Sources

Evidence basis: Recorded source evidence
Setup runtime verification: Not independently runtime-tested by MCP100 for this setup workflow.

Reviewed Oct 5, 2026 · Latest source observation Oct 8, 2026

Package registry metadata ↗
@testivai/mcp
Recorded package identity, distribution or download evidence.
GitHub repository ↗
testivai/testivai-oss
Repository reference; implementation attribution is not established for every document.
Includes references whose attribution is not established.

Setup · Watchlist

Testing / Debugging setup

Publisher instructions.

Source evidence observed Oct 14, 2026.

Additional setup details required — see the specific gaps below.

Set up testing / debugging integration
Publisher instructions
Testing tool; check source prerequisites and avoid running against production target as smoke test.
publisher documented install or connect · Publisher instructions
claude mcp add testivai -- npx -y @testivai/mcp
Configure / authenticate
MCP100 setup guide
Visual test result inspection is read-oriented; snapshot approvals mutate testing baselines and are not installation smoke tests.
Check connection
Suggested check
User-side suggested check (not executed): inspect registered tools/connection and perform publisher-supported read-only status/list operations. Do not send, purchase, publish, delete, sign, trade or alter data as a smoke test.
Remove / uninstall / revoke
MCP100 setup guide
Generic client guidance (not publisher-specific uninstall): unregister MCP entry in the chosen client. If installed globally, remove the npm package separately. Revoke provider access separately through account controls, where supported. This does not erase local data or revoke funds/credentials automatically.
Unregister the named MCP entry from the actual host separately from removing a global npm install; no claim that either revokes remote credentials.
Remove local stored keys/tokens and revoke or rotate corresponding provider authorization using documented provider account controls if available; exact procedure not verified for this package. Wallet private keys cannot be revoked.

Official Documentation & Sources

What it does

MCP server in the TestivAI OSS suite. It exposes visual test results and diff images so the same coding agent that changed a UI can inspect regressions locally. Baseline approval is intentionally a human decision.

Capability: Security & Testing
Resource type: Testing / Debugging
Best fit: Local-first visual regression evidence and diff inspection for AI coding agents.

Representative uses

Read visual regression results
Inspect diff images and DOM/style evidence
Explain changed snapshots
Support Playwright/Selenium/WebdriverIO workflows

Limitations

Baseline approval must remain human-controlled.
Dynamic content can create false diffs unless ignored/tolerated.
Exact MCP tool inventory can evolve with the suite.
No hosted AI service is needed for local diffing.
Do not auto-approve visual baselines; project explicitly keeps approval human-controlled.

Reviewed access

Access & risk

Review the exact access boundary before connecting this MCP to accounts, credentials, or sensitive data.

Authentication: Not documented in the reviewed sources
Permissions: Read local visual-test evidence; approval should not be automated by agent.
Data handling: Visual reports/diffs stay local in the OSS workflow.
Cost: free_open_source
Risk and cautions: Primarily read/verification; biggest safety rule is not allowing the agent to self-approve baselines.

Access scope and cautions

Write actions: None in the reviewed inventory
Sensitive access: Screenshots/DOM snapshots may contain application data
Risk level: low

FAQ

Based on reviewed evidence as of 2026-10-08. Watchlist tracking is not a recommendation or a security certification.

Which tasks are documented for this implementation?

The reviewed tasks include: Read visual regression results; Inspect diff images and DOM/style evidence; Explain changed snapshots; Support Playwright/Selenium/WebdriverIO workflows.

What setup and authentication are documented?

Recorded installation/entrypoint step: npx -y @testivai/mcp. Recorded transport: stdio. Prerequisites: Node 20+; TestivAI visual test output. These are the documented setup details, not a claim that installation or client compatibility was independently tested. Authentication: No API key for local-first OSS use.

Can it modify data or perform write operations?

The reviewed inventory lists no write actions. An empty inventory alone does not establish read-only authority. Recorded permission boundary: Read local visual-test evidence; approval should not be automated by agent.

What data-handling boundary should I consider?

Visual reports/diffs stay local in the OSS workflow. Recorded sensitive-access context: Screenshots/DOM snapshots may contain application data.

What limitations and risks should I check before use?

Primarily read/verification; biggest safety rule is not allowing the agent to self-approve baselines. Baseline approval must remain human-controlled. Dynamic content can create false diffs unless ignored/tolerated. Exact MCP tool inventory can evolve with the suite. No hosted AI service is needed for local diffing. Do not auto-approve visual baselines; project explicitly keeps approval human-controlled. Runtime behavior, dependencies and undisclosed vulnerabilities were not independently audited. Documented tool authority and any remote data handling must be checked before installation or use.