whitehat-swarm-mission-control.burnersforwinter.chatgpt.site

# Passive security review — DeepHat | Autonomous security swarm

Run: f58cecad-9eef-4ffc-99b6-3aea341d6acc  
Status: **completed**  
Reviewed: `whitehat-swarm-mission-control.burnersforwinter.chatgpt.site`, `www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site`  
Started: 2026-10-25T22:50:12.549Z  
Requests: 18 of 40 budgeted, at most 1/s  
Discovered pages: 5
Tier: 1

## Summary

**6** configuration issues: 1 high, 3 medium, 2 low.

### Pages fetched

- https://whitehat-swarm-mission-control.burnersforwinter.chatgpt.site/api-access/privacy → HTTP 200
- https://whitehat-swarm-mission-control.burnersforwinter.chatgpt.site/api-access/terms → HTTP 200
- https://whitehat-swarm-mission-control.burnersforwinter.chatgpt.site/ → HTTP 200
- https://whitehat-swarm-mission-control.burnersforwinter.chatgpt.site/contact → HTTP 404
- https://whitehat-swarm-mission-control.burnersforwinter.chatgpt.site/contact-us → HTTP 404

## Checks

| Check | Host | Severity | Result |
| --- | --- | --- | --- |
| The site serves HTTPS | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | High | fail |
| HTTP Strict Transport Security | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Medium | fail |
| Content Security Policy | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Medium | fail |
| MIME type sniffing is disabled | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Low | fail |
| Framing is restricted | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Medium | fail |
| Referrer policy | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Low | fail |
| Cookie attributes | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Info | pass |
| Server version disclosure | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Info | pass |
| Cross-origin resource sharing | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Info | pass |
| HTTP redirects to HTTPS | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Info | pass |
| Git metadata is not exposed | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Info | pass |
| Environment file is not exposed | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Info | pass |
| Server status page is not exposed | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Info | pass |
| Exposure probe | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site, www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Low | error |
| TLS certificate | whitehat-swarm-mission-control.burnersforwinter.chatgpt.site | Info | pass |

## Findings

### 1. The site serves HTTPS (High)

An HTTPS request to www.whitehat-swarm-mission-control.burnersforwinter.chatgpt.site failed: write EPROTO 80FDC76A2D7F0000:error:0A000410:SSL routines:ssl3_read_bytes:ssl/tls alert handshake failure:../ssl/record/rec_layer_s3.c:918:SSL alert number 40


**Suggested fix.** Serve the site over HTTPS with a valid certificate and redirect HTTP traffic.

### 2. HTTP Strict Transport Security (Medium)

No Strict-Transport-Security header was returned.

**Suggested fix.** Add Strict-Transport-Security: max-age=31536000; includeSubDomains. Start with a short max-age while testing.

### 3. Content Security Policy (Medium)

No Content-Security-Policy header was returned.

**Suggested fix.** Start with Content-Security-Policy-Report-Only. Inventory external scripts and inline code, then test a site-specific policy before enforcing it.

### 4. Framing is restricted (Medium)

Neither X-Frame-Options nor a CSP frame-ancestors directive restricts framing.

**Suggested fix.** Send X-Frame-Options: SAMEORIGIN or a CSP frame-ancestors directive.

### 5. MIME type sniffing is disabled (Low)

No X-Content-Type-Options header was returned.

**Suggested fix.** Add X-Content-Type-Options: nosniff.

### 6. Referrer policy (Low)

No Referrer-Policy header was returned.

**Suggested fix.** Add Referrer-Policy: strict-origin-when-cross-origin.

## Patch

Configuration suggestions for staging review; see `patch/README.md` for placement and verification.

### _headers

```
# Review in staging. CSP is report-only and does not remediate the CSP finding.
/*
  Strict-Transport-Security: max-age=31536000; includeSubDomains
  Content-Security-Policy-Report-Only: default-src 'self'; object-src 'none'; base-uri 'self'; frame-ancestors 'none'; script-src 'self'
  X-Content-Type-Options: nosniff
  X-Frame-Options: SAMEORIGIN
  Referrer-Policy: strict-origin-when-cross-origin
```

### nginx.conf

```
# Review in staging. CSP is report-only; tailor it before enforcement.
# Add inside the server { } block, then reload nginx.
add_header Strict-Transport-Security "max-age=31536000; includeSubDomains" always;
add_header Content-Security-Policy-Report-Only "default-src 'self'; object-src 'none'; base-uri 'self'; frame-ancestors 'none'; script-src 'self'" always;
add_header X-Content-Type-Options "nosniff" always;
add_header X-Frame-Options "SAMEORIGIN" always;
add_header Referrer-Policy "strict-origin-when-cross-origin" always;
```

### apache.conf

```
# Review in staging. CSP is report-only; tailor it before enforcement.
# Add inside the <VirtualHost> block or .htaccess, then reload Apache.
Header always set Strict-Transport-Security "max-age=31536000; includeSubDomains"
Header always set Content-Security-Policy-Report-Only "default-src 'self'; object-src 'none'; base-uri 'self'; frame-ancestors 'none'; script-src 'self'"
Header always set X-Content-Type-Options "nosniff"
Header always set X-Frame-Options "SAMEORIGIN"
Header always set Referrer-Policy "strict-origin-when-cross-origin"
```

## Requires site-specific review

- https-availability
- csp

## Application findings (deep pass)

### 1. Unauthenticated exposure of internal swarm execution telemetry, credentials, and full reports via /api/swarm/state (Medium)

Location: /api/swarm/state

The public "spectator" endpoint `GET /api/swarm/state` returns the complete internal recorded-run payload to any unauthenticated visitor. It includes the swarm's raw shell commands, credential values typed into the login form, internal tooling paths, agent identifiers, the full penetration-test reports and patch for the internal DeskFlow fixture, internal budget/inference accounting, and the operator policy configuration. Only the operator control endpoints (`/api/operator/*`) and `/operator` are access-controlled; the corresponding read model at `/api/swarm/state` is not.

Evidence: Anonymous request and response (headers trimmed):

```http
GET /api/swarm/state HTTP/1.1
Host: whitehat-swarm-mission-control.burnersforwinter.chatgpt.site

HTTP/1.1 200 OK
Content-Type: application/json
Cache-Control: no-store
Server: cloudflare
```

Body excerpts (32,568 bytes total):

```json
{"backendRevision":"spectator-proof-6","job":{"status":"passed","attempts":3,...},
 "result":{"findingCount":3,"verifiedFixCount":3,
   "report":"# Security Penetration Test Report ... **Key findings** ... Privilege escalation and tenant relocation (Critical) ...",
   "patch":"--- a/app.py\n+++ b/app.py\n@@ -85,17 +85,21 @@ ...",
   "findings":[{"id":"vuln-0002","title":"Privilege escalation and cross-tenant account relocation via mass assignment on PATCH /api/profile","severity":"critical"}],
   "scope":"Owned synthetic multi-company DeskFlow portal on loopback only"},
 "telemetry":{"terminal":[{"command":"... agent-browser --session swarm fill @e4 \"deskflow-demo-only\" ... open http://127.0.0.1:8765/login","output":"..."}],
   "source":"native-strix-sqlite"},
 "inference":{"limit":2,"remaining":1.79,"maxRequests":600,"inputTokens":10470602},
 "policy":{"scope":"Built-in synthetic DeskFlow portal only","hosting":"Cloudflare Containers","model":"deepseek-ai/DeepSeek-V4.1-Flash-1017","treasury":"Not connected"}}
```

Control — the authenticated sibling routes correctly reject the same anonymous caller:

```json
GET  /api/operator/state  -> 403 {"error":"Operator sign-in required."}
POST /api/queue/next      -> 403 {"error":"Operator sign-in required."}
GET  /api/swarm/state     -> 200 (full internal document)
```



## Basis and limits

The baseline is passive: GET and HEAD against public addresses only. DNS is resolved once and connections are pinned; the site stays inside its own registrable domain; redirects are reported, never followed; the rate and request budget are enforced.

Checks cover transport, response headers, cookie flags, version banners, CORS, the certificate, and a few well-known exposure paths.

## What this review did not do

No exploitation, login, form submission, credential use, file upload, or application-logic testing in the passive baseline. No finding is confirmed exploitable; exposures are recorded as directly observed. Nothing was changed on the server.

## Delivery

`owner-email.md` and `github-issue.md` are generated from these findings. Nothing is sent unless the `send` command runs.