8d757a99dd
Intensive end-to-end testing (a live gateway-MCP <-> agent integration harness +
an adversarial redaction/abuse probe + an independent security audit) surfaced
three real issues in the shipped read-only diagnostic agent. All run in lib/**,
which is packaged in the app.
1. grep ReDoS froze the Electron main process. read_log compiled the
client-supplied grep into `new RegExp(grep, 'i')` and ran it synchronously over
the log tail IN the main process. A catastrophic pattern (e.g. "(a+)+$" against
a long line) hangs the whole app — empirically confirmed (8s timeout, killed).
JS regex is synchronous and uncancellable, so grep is now a case-insensitive
literal substring filter with "|" alternation ("error|timeout|502"). Provably
linear-time; covers the real diagnostic need.
2. Prototype-chain whitelist bypass. The op table was a plain object literal, so
handle("constructor" | "toString" | "valueOf", ...) resolved an inherited
Object.prototype function, passed the `typeof fn === 'function'` guard and
returned {ok:true}. Harmless functions today, but a whitelist-integrity hole.
Now guarded with a string check + Object.prototype.hasOwnProperty.
3. Redaction defense-in-depth gaps. redactLogText now also scrubs: basic-auth URL
passwords (scheme://user:pass@host), Authorization: Basic, JWTs (eyJ...x.y.z),
and bare/JSON session= values. Mostly theoretical in today's readable logs
(secret-bearing bodies go to the excluded doodstream-debug.log; other hosters
throw static strings) but matters as the verbose-logging surface grows.
Verified: 383 app tests (incl. new regression tests for all three), the live
gateway-MCP integration harness (all 14 tools, zero leaks, error paths), the
adversarial probe (14/14+ secret shapes scrubbed, ReDoS 1ms, lockout, malformed
args), e2e gate, lint 0 errors. Only residual: a standalone high-entropy blob with
zero key/Bearer/URL context — inherent to any denylist, acknowledged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
||
|---|---|---|
| .. | ||
| test | ||
| verify | ||
| .gitignore | ||
| agent-client.js | ||
| code.js | ||
| index.js | ||
| package-lock.json | ||
| package.json | ||
| README.md | ||
| registry.example.json | ||
| registry.js | ||
mhu-diagnostics-gateway
A standalone local stdio MCP gateway for remote, read-only diagnostics of the Multi-Hoster-Uploader app.
It is two things at once:
- an MCP server to Claude Code (stdio transport), exposing read-only diagnostic tools, and
- a plain WebSocket client to a diagnostic agent running inside the Electron app on a remote Windows server.
The operator enables "Diagnose-Zugriff" on a server, copies the connection code, and tells
Claude server <name> at <host>, code <CODE>. Claude calls connect_server(code, host) and then
the read-only diagnostic tools. After the first successful connect the server is remembered under
its label, so later you can just say connect_server(label:"prod-3") with no code.
This package is fully self-contained. It does not import anything from the parent Electron app and is not part of the app build.
Install
cd gateway
npm install
Requires Node >= 18.
Register with Claude Code (one time)
claude mcp add --transport stdio mhu-diag -- node "C:\Users\ploet\Desktop\Claude Projekte\multi-hoster-uploader\gateway\index.js"
Adjust the absolute path if you cloned the repo elsewhere.
Usage
In Claude Code, tell Claude:
server prod-3 at 127.0.0.1, code mhu1_<...>
Claude will call connect_server and then diagnostic tools such as server_health
(the one-shot "what's wrong" hub), read_log, list_errors, get_queue_state,
get_rotation_state, and so on.
Security
- Read-only. No screen access, no input injection, no writes. Only reads logs, errors, queue/history/config (redacted), rotation and system info.
- The code is a secret — it carries the auth token. Do not paste it anywhere public.
- The safe default is to reach the agent over
127.0.0.1via an SSH local port-forward or WireGuard. Seedocs/remote-diagnostics-setup.md. registry.jsonstores tokens and is git-ignored — never commit it.