Atlas / MCP servers / cliffhall / Puzzlebox

PuzzleboxBLOCK

mcp/cliffhall/puzzlebox

An MCP server that hosts finite state machines as dynamic resources that multiple clients can subscribe to and be updated when their state changes.

Verdict
BLOCK
Grade
D
Trust score
69 /100
Exposed tools
5 4r · 1w · 0d
Transport
sse · streamable-http
License
MIT
Stars
32
01

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

Coordinating agents with state machines

An MCP server that hosts finite state machines as dynamic resources that clients can subscribe to and be updated when their state changes.

Feature Roadmap and Status

A work in progress. Much is done, a few bits remain to be done.

  • [x] Unit tests for tool code
  • [x] Integration tests for MCP server (both SSE and StreamableHttp)
  • [x] SSE transport
  • [x] StreamableHttp transport
  • [x] Multiple simultaneous client connections (both transports)
  • [x] Create puzzles (state machines) as dynamic resources
  • [x] Subscribe to puzzles
  • [x] Receive update notifications when puzzles change
  • [x] Get puzzle snapshot (current state and available actions)
  • [x] Change puzzle state by performing action on puzzle
  • [ ] Resource creation by puzzle state (used by agents and guard prompts)
  • [ ] Transition guard prompt creation (uses resources by state)
  • [ ] Transition guard via sampling
  • [ ] Command line REPL
  • [ ] Demo via REPL

What problem does puzzlebox address?

Collaboration and coordination are related but different problems. Collaboration is good for non-trivial yet relatively simple tasks. To tackle long horizon efforts, puzzlebox solves for coordination.

Teams need coordination

Marshalling multiple agents toward a big goal is tougher than just breaking down a request into tasks, assigning them to available agents and enabling collaboration between them.

Just as a few agents can collaborate to complete a small project, several teams of process-aware agents need to operate within distinct project phases to tackle long horizon efforts.

Consider enterprise-level software development processes:

  • A large software project typically moves th
Read from source at commit 17e2327d49ffOBSERVED · 2026-10-09
02

Connect

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.

claude-code
claude mcp add puzzlebox -- npx -y [email protected]
claude-desktop
{
  "mcpServers": {
    "puzzlebox": {
      "command": "npx",
      "args": [
        "-y",
        "[email protected]"
      ]
    }
  }
}
03

Exposed tools (5)

4 read · 1 write · 0 destructive.

ToolRiskDescription
add_puzzlewriteAdd a puzzle
calculatorreadA calculator
count_puzzlesreadCount puzzles
get_puzzle_snapshotreadGet a snapshot of a puzzle (its current state and available actions).
perform_action_on_puzzlereadPerform an action on a puzzle (attempt a state transition).
04

Trust audit

BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.

LayerWhat it checksResult
L0Provenance & inventoryPASS
L1Static analysis of the codeFAIL
L2Instruction surface (what it tells the agent)PASS
L3Class-specific surfaceWARN
L4Behavioural (sandbox)SKIPPED

What the source does

Filesystem
none-observed
Network
declared (5 observation(s))
Shell
declared (1 observation(s))
Dependencies
not all pinned
Secrets in source
none-found

Findings (13)

HIGHCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95
dist/repl.js:2220
var makeValidate = new Function(
Why it matters. evaluates text as code
Fix. remove; use a parser or a dispatch table
MEDIUMAuth / authz · mcp.remote_no_auth · CWE-287, CWE-862
sse,streamable-http
Why it matters. a network transport with no auth environment variable found
Fix. require a token
LOWInventory / provenance · inv.hidden_file · CWE-1104
.prettierignore
.prettierignore
Why it matters. hidden member outside the usual dotfiles
Fix. review its purpose
LOWFilesystem / path · fs.traversal · CWE-22, CWE-59
src/stores/__tests__/Puzzle.test.ts:2
import { getTestPuzzleConfig } from "../../common/utils.ts";
LOWFilesystem / path · fs.traversal · CWE-22, CWE-59
src/stores/__tests__/PuzzleStore.test.ts:2
import { getTestPuzzleConfig } from "../../common/utils.ts";
LOWFilesystem / path · fs.traversal · CWE-22, CWE-59
src/tools/__tests__/puzzles.test.ts:2
import { getTestPuzzleConfig } from "../../common/utils.ts";
LOWFilesystem / path · fs.traversal · CWE-22, CWE-59
src/tools/__tests__/puzzles.test.ts:4
import { PuzzleConfig } from "../../common/schemas.js";
LOWFilesystem / path · fs.traversal · CWE-22, CWE-59
src/tools/__tests__/puzzles.test.ts:5
import PuzzleStore from "../../stores/PuzzleStore.ts";
LOWObfuscation / stealth · obf.decode_call · CWE-506, CWE-94
dist/repl.js:7123
const decoded = JSON.parse(atob(base64));
LOWObfuscation / stealth · obf.decode_call · CWE-506, CWE-94
src/puzzlebox.ts:181
const decodedCursor = parseInt(atob(cursor), 10);
LOWSupply chain · supply.unpinned · CWE-829, CWE-1357
package.json
@modelcontextprotocol/inspector, @modelcontextprotocol/sdk, @types/supertest, mcp-proxy, node-fetch, tsx, supertest, @types/cors
Why it matters. 19 dependency range(s) float
Fix. pin exact versions or ship a lockfile
INFOInventory / provenance · inv.oversize · CWE-1104
dist/sse.js
dist/sse.js
Why it matters. 2365839 bytes not read
INFOInventory / provenance · inv.oversize · CWE-1104
dist/streamableHttp.js
dist/streamableHttp.js
Why it matters. 2380668 bytes not read

Gates applied: no_behavioural_pass.

Audited 2026-10-09 · audit v0.4.1 · source sha 17e2327d49fffull audit observations/trust-audit/mcp-server/cliffhall__puzzlebox.json · Report an issue / request a re-scan
05

Audit history

Every audit this server has had. A grade with a past is a grade somebody is still checking.

DateSourceVerdictGradeScoreChange
2026-10-0917e2327d49ffBLOCKD69first audit
06

Questions

What is the Puzzlebox MCP server?

An MCP server that hosts finite state machines as dynamic resources that multiple clients can subscribe to and be updated when their state changes.

What tools does Puzzlebox expose?

5 in total: 4 read-only, 1 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.

Is Puzzlebox safe to connect to an agent?

No — not without reading the findings first. The audit graded it D (69/100) and found 1 critical or high issue in the source. Each one is listed on this page with the file and line it is on.

What credentials does Puzzlebox need?

No credential environment variables were found in its source, so it appears to need none.

How does Puzzlebox run?

It speaks sse and streamable-http, so it runs as a service you connect to over the network. It is published on npm as puzzlebox at 1.0.0.

How current is this page?

The grade is for one exact copy of the source (17e2327d49ff), read on 2026-10-09. The repository is watched and re-audited when it changes.

Advertisement