MCP server · Audited from source
whitenightshadow-camoufox-reverse MCP server
by whitenightshadow/camoufox-reverse-mcpMIT licenceAudited 12 Nov 2026 · source d05f2692dfea
Is it safe? With care — CAUTION, grade B (83/100). It exposes 42 tools: 32 read, 8 write and 2 destructive. The audit recorded 15 findings (0 critical, 0 high, 5 medium, 10 low). Maintainer's description: 🔧 反检测浏览器 MCP Server | 39 个 JS 逆向工具 + C++ 引擎层 JSVMP 指纹追踪
Cite
Cite this page
ModelsAtlas. “whitenightshadow-camoufox-reverse MCP server (whitenightshadow/camoufox-reverse-mcp)”. Retrieved 12 Nov 2026. https://modelsatlas.com/mcp/whitenightshadow/camoufox-reverse-mcp
Trust audit
Audit 0.4.1 · full source read 12 Nov 2026
15 findings, listed below with the file and line each was found on. Grade A needs a sandboxed behavioural run too; this audit read the source only.
What the source does
- FilesystemUndeclared3 observations not in its description
- NetworkDeclared6 observations
- ShellDeclared1 observation
- DependenciesPinnedNothing floats
- SecretsNone foundNo keys in source
“Declared” means the server says it does this and the source agrees. Browser, file and shell access are expected of some kinds of server.
Five audit layers
| Layer | What we check | Result |
|---|---|---|
L0 Provenance | Repo owner, licence, archive status, activity | PASS |
L1 Static source | Secrets, path traversal, supply chain | WARN 14 findings |
L2 Instructions | Prompt injection and authority framing in docs and tool descriptions | PASS |
L3 Inventory | Declared tools match the source | WARN 1 finding |
L4 Sandbox | Run in isolation and watch behaviour | SKIPPED |
Provenance & runtime
Who published it, and what the audit could and could not run
Provenance
- Repository owner
whitenightshadowUser accountAccount type recorded when we read the repository. - PackagePublished on PyPI
camoufox-reverse-mcp, no version recorded - LicenceMITFrom the repository, as read at the audited commit.
- Last pushTodayActivity of the repository when we read it.
- Source commit readd05f2692dfeaThe audit covers exactly this commit, read on 12 Nov 2026.
Runtime
- Sandbox runNot run yetThis audit read the source only; it did not execute the server.
- Tool inventory42 declaredRead from the source and classed read, write or destructive.
Every value here is read from the record's own repository and audit.Suggest a correction
Findings (15)
Sorted by severity
MEDIUMCode injectioncode.dynamic_import · src/camoufox_reverse_mcp/tools/environment.py:43
m = importlib.import_module(dep)
- Why
- No further detail recorded for this rule.
- Fix
- —
MEDIUMFilesystem / pathfs.destructive · src/camoufox_reverse_mcp/browser.py:250
shutil.rmtree(trace_base_dir, ignore_errors=True)
- Why
- No further detail recorded for this rule.
- Fix
- —
MEDIUMFilesystem / pathfs.destructive · src/camoufox_reverse_mcp/property_trace.py:479
shutil.rmtree(run)
- Why
- No further detail recorded for this rule.
- Fix
- —
MEDIUMNetwork egressnet.raw_ip · src/camoufox_reverse_mcp/__main__.py:22
parser.add_argument("--proxy", type=str, help="Proxy server URL (e.g. http://127.0.0.1:7890)")- Why
- No further detail recorded for this rule.
- Fix
- —
MEDIUMFilesystem / pathmcp.destructive_tools
remove_hooks, reset_browser_state
- Why
- 2 tool(s) can delete or overwrite
- Fix
- prefer a read-only mode or scoped tokens; the page states the blast radius
LOWCode injectioncode.eval_exec · tests/test_env_capture.py:16
check = "const f = eval(require('fs').readFileSync(0, 'utf8')); process.stdout.write(typeof f);"- Why
- evaluates text as code
- Fix
- remove; use a parser or a dispatch table
LOWCode injectioncode.eval_exec · tests/test_stage_taps.py:103
src = 'var gen = new Function("l", "' + STAGE_BODY + '");\n' + STAGE_CALL- Why
- evaluates text as code
- Fix
- remove; use a parser or a dispatch table
LOWInsecure cryptocrypto.weak_hash · tests/test_collection.py:223
result = await verification.verify_signer_offline("async ({text}) => ({sign: require('node:crypto').createHash('md5').update(text).digest('hex')})", [{'input':{'text':'abc'},'expected':{'sign':'900150- Why
- No further detail recorded for this rule.
- Fix
- —
LOWNetwork egressnet.raw_ip · README.md:156
"args": ["-m", "camoufox_reverse_mcp", "--headless", "--proxy", "http://127.0.0.1:7890"]
- Why
- No further detail recorded for this rule.
- Fix
- —
LOWNetwork egressnet.raw_ip · README.md:162
| `--proxy <url>` | 代理地址,例如 `http://127.0.0.1:7890` |
- Why
- No further detail recorded for this rule.
- Fix
- —
LOWNetwork egressnet.raw_ip · README_en.md:156
"args": ["-m", "camoufox_reverse_mcp", "--headless", "--proxy", "http://127.0.0.1:7890"]
- Why
- No further detail recorded for this rule.
- Fix
- —
LOWNetwork egressnet.raw_ip · README_en.md:162
| `--proxy <url>` | Proxy URL, for example `http://127.0.0.1:7890` |
- Why
- No further detail recorded for this rule.
- Fix
- —
LOWObfuscation / stealthobf.decode_call · src/camoufox_reverse_mcp/hooks/crypto_hook.js:31
__mcpDefine(window, 'atob', __mcpNative(function atob(s) {- Why
- No further detail recorded for this rule.
- Fix
- —
LOWObfuscation / stealthobf.decode_call · tests/test_network_evidence.py:63
@pytest.mark.parametrize('raw',[b'',bytes(range(256)), '字节一致的 JS\n'.encode(), bytes.fromhex('0061736d01000000')])- Why
- No further detail recorded for this rule.
- Fix
- —
LOWObfuscation / stealthobf.zero_width · tests/test_stage_break.py:120
return json.loads(out.stdout.lstrip(""))- Why
- No further detail recorded for this rule.
- Fix
- —
Tools it exposes
42 declared in source · 2 destructive
2 tools can delete or overwrite. An agent that can be talked into calling a tool can be talked into calling this one, so give it a token scoped to what you need.
| Tool | Risk | What it does |
|---|---|---|
remove_hooks | destructive | Remove installed hooks and restore original objects in-place. |
reset_browser_state | destructive | Reset browser residual state after serializing trace transitions. |
check_environment | write | One-stop self-check of MCP environment, dependencies, and browser state. |
evaluate_js | write | Execute an arbitrary JavaScript expression in the page context and return the result. |
hook_jsvmp_interpreter | write | Install a JSVMP runtime probe in the page main world. |
import_state | write | Import browser state from a JSON file by creating a new context. |
list_trace_files | write | List all trace files on disk (for post-hoc analysis). |
query_trace_file | write | Query a specific historical trace file (post-hoc analysis). |
save_response_body | write | Save captured response bytes (JS/WASM/JSON/binary) without refetch/replay. |
trace_property_access | write | Control/query the current Gecko native PropertyTracer run. |
analyze_cookie_sources | read | Attribute every observed cookie to its source (HTTP header vs JS). |
click | read | Click on a page element. |
close_browser | read | Close the browser after serializing any native trace transition. |
compare_env | read | Collect the browser environment for comparison with Node.js/jsdom. |
compare_network_requests | read | Compare 2..10 captured requests without issuing requests or launching a browser. |
cookies | read | Cookie management (v0.9.0 unified). |
export_network_capture | read | Export a versioned JSON snapshot of this manager |
export_state | read | Export the complete browser state (cookies + storage) to a JSON file. |
get_console_logs | read | Get console output collected from the page. |
get_network_request | read | Get full details of a specific captured network request. |
get_page_info | read | Get page metadata plus a zero-based frame list for targeted tools. |
get_request_initiator | read | Get the JS call stack that initiated a network request. |
get_storage | read | Get the contents of localStorage or sessionStorage. |
get_trace_data | read | Retrieve in-page traces and BrowserManager |
hook_function | read | Hook or trace a function (v0.9.0 unified). |
inject_hook_preset | read | Inject a pre-built hook into the page main world. |
instrumentation | read | JSVMP source-level instrumentation (v0.9.0 unified). |
intercept_request | read | Intercept network requests matching a pattern. |
launch_browser | read | Launch the Camoufox anti-detection browser, or attach to a running one. |
list_network_requests | read | List captured network requests with optional filters. |
navigate | read | Navigate to a URL, with optional hook pre-injection and redirect tracing. |
network_capture | read | Unified network capture control (v0.9.0). |
reload | read | Reload the current page, preserving any init scripts. |
scripts | read | Script inspection (v0.9.0 unified). |
search_code | read | Search keyword in loaded scripts (v0.9.0 unified). |
stage_break | read | Per-state deep capture for `instrumentation(tap_stage_calls=True)`. |
take_screenshot | read | Take a screenshot of the current page or a specific element. |
take_snapshot | read | Get a bounded accessibility tree, or a labelled DOM fallback. |
trace_session | read | Record a Playwright trace of the current browser context. |
type_text | read | Type text into an input field with realistic keystroke delays. |
verify_signer_offline | read | Verify a signer against explicit expected values without sending requests. |
wait_for | read | Wait for an element to appear or a network request matching a URL pattern. |
Install
audited commit d05f269
Read the findings first. The most severe is Code injection (medium) at src/camoufox_reverse_mcp/tools/environment.py.
One-click links open your editor with the config below filled in, exactly as the snippets show it. The audit covers commit d05f269 only.
claude mcp add camoufox-reverse-mcp -- uvx camoufox-reverse-mcp
{
"mcpServers": {
"camoufox-reverse-mcp": {
"command": "uvx",
"args": [
"camoufox-reverse-mcp"
]
}
}
}The command does not pin a version; the audit covers commit d05f2692dfea. Transport: stdio. No environment variables required.
Audit history
- !CAUTION B 83 / 100
First audit ·
d05f2692dfea
README badge
For maintainers · updates on every re-audit
[](https://modelsatlas.com/mcp/whitenightshadow/camoufox-reverse-mcp/)
The badge always shows the latest audit; a maintainer cannot pin an old verdict.
Other ratings
Cited, never blended into our score
No third-party security ratings are recorded for this server yet. When they exist they appear here with their source and date, beside, not inside, our grade.
Questions about whitenightshadow-camoufox-reverse MCP server
What is the whitenightshadow-camoufox-reverse MCP server?
What tools does whitenightshadow-camoufox-reverse expose?
Is whitenightshadow-camoufox-reverse safe to connect to an agent?
What credentials does whitenightshadow-camoufox-reverse need?
How does whitenightshadow-camoufox-reverse run?
How current is this page?
Sources
- whitenightshadow/camoufox-reverse-mcp on GitHub — the source we read, its licence, stars and activityObserved · 12 Nov 2026
- ModelsAtlas trust audit — CAUTION verdict for commit d05f2692dfeaObserved · 12 Nov 2026
Spotted something wrong? Report a correction · Methodology
Keep going in your AI
Opens with this page as the source. We add nothing else to the prompt.
- What is the whitenightshadow-camoufox-reverse MCP server?
- What tools does whitenightshadow-camoufox-reverse expose?
- Is whitenightshadow-camoufox-reverse safe to connect to an agent?