Get GatherBLOCK
Exoskeleton for autonomous assistant
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
Remote Browser is an open-source, self-hostable tool for every autonomous assistant. It creates isolated, sandboxed browser environments for every user, always on and accessible 24/7 from anywhere at any time, delivering robust privacy, strict access control, and a complete audit trail.
For the detailed documentation, visit docs.remotebrowser.ai.
Working through its harness, an LLM (large language model) can remotely control these browsers to automate complex workflows: preparing employee onboardings, managing expenses, generating sales reports, conducting market research, and many more.
Remote Browser works with the common web automation tools (Playwright, Puppeteer, etc), with autonomous coding assistants (Claude Code, ChatGPT Codex, Cursor, Antigravity, OpenCode, etc), and with the new generation of intelligent assistants (OpenClaw, Hermes, Pi, etc).
[](https://youtu.be/w3pg9wJhfqY)
9e68504fb9c6OBSERVED · 2026-10-08Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add remotebrowser --env SMTP_PASSWORD=${SMTP_PASSWORD} -- npx -y [email protected]{
"mcpServers": {
"remotebrowser": {
"command": "npx",
"args": [
"-y",
"[email protected]"
],
"env": {
"SMTP_PASSWORD": "${SMTP_PASSWORD}"
}
}
}
}Exposed tools (1)
1 read · 0 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
calm-otter | read | For QA testing |
Trust audit
BLOCKgrade F · trust 53/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (5 observation(s))
- Network
- declared (11 observation(s))
- Shell
- declared (10 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- found
Findings (23)
const { code, stdout } = await exec(['port', name, String(containerPort)]);const { code, stdout, stderr } = await exec(args);const { code } = await exec(dialect.existsArgs(containerName(browserId)));const { code, stdout } = await exec(['inspect', '--format', '{{.State.Running}}', containerName(browserId)]);const { code, stderr } = await exec(['kill', name]);databaseUrl && isProduction && env.DATABASE_SSL_MODE !== 'disable' ? { rejectUnauthorized: false } : undefined;assert.deepEqual(ssl, { rejectUnauthorized: false });const MOCK_BROWSERFLEET_URL = `http://127.0.0.1:${MOCK_BROWSERFLEET_PORT}`;const res = await fetch(`http://127.0.0.1:${relay.port}/anything`);process.env.BROWSERFLEET_URL = `http://127.0.0.1:${cdpStub.address().port}`;process.env.BROWSERFLEET_URL = `http://127.0.0.1:${cdpStub.address().port}`;--env DATABASE_URL=postgres://remotebrowser:[email protected]:5432/remotebrowser \
.oxfmtrc.json
.oxlintrc.json
'../../Collaborators/[email protected]',
for (const browserId of ['x/../../admin', '..', 'a.b', 'a%2f']) {for (const internalBrowserId of ['x/../../admin', '..', 'a.b', 'a%2f', '']) {import { eta } from '../../render.js';import { createCsrfToken } from '../../auth/csrf.js';if curl -fsS http://127.0.0.1:3000/health; then
@hono/node-server, @opentelemetry/api, @opentelemetry/api-logs, @opentelemetry/exporter-logs-otlp-proto, @opentelemetry/exporter-trace-otlp-proto, @opentelemetry/resources, @opentelemetry/sdk-logs, @o
<iframe width="560" height="315" src="https://www.youtube.com/embed/w3pg9wJhfqY" title="Remote Browser demo" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope
The app does not use passwords. Instead, it signs users in with an email link (also called a "magic link") to verify their identity. When a user requests a sign-in link, the app stores a time-limited,
Gates applied: no_behavioural_pass.
9e68504fb9c6full audit observations/trust-audit/mcp-server/mcp-getgather__get-gather.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-08 | 9e68504fb9c6 | BLOCK | F | 53 | first audit |
Questions
What is the Get Gather MCP server?
Exoskeleton for autonomous assistant
What tools does Get Gather expose?
1 in total: 1 read-only, 0 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is Get Gather safe to connect to an agent?
No — not without reading the findings first. The audit graded it F (53/100) and found 7 critical or high issues in the source. Each one is listed on this page with the file and line it is on.
What credentials does Get Gather need?
It reads SMTP_PASSWORD from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How current is this page?
The grade is for one exact copy of the source (9e68504fb9c6), read on 2026-10-08. The repository is watched and re-audited when it changes.