# Prototype Pollution

> Detect prototype pollution via object merge/clone/assign operations where __proto__ or constructor.prototype keys can modify Object.prototype.

- **Type:** Skill
- **Install:** `agentstack add skill-byamb4-find-cve-agent-prototype-pollution`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [ByamB4](https://agentstack.voostack.com/s/byamb4)
- **Installs:** 0
- **Category:** [Security](https://agentstack.voostack.com/c/security)
- **Latest version:** 0.1.0
- **License:** Apache-2.0
- **Upstream author:** [ByamB4](https://github.com/ByamB4)
- **Source:** https://github.com/ByamB4/find-cve-agent/tree/main/skills/prototype-pollution

## Install

```sh
agentstack add skill-byamb4-find-cve-agent-prototype-pollution
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Prototype Pollution Detection

## When to Use

Audit merge/clone/deep-assign utilities, query string parsers, JSON parsers, config mergers, and any package that recursively sets object properties from untrusted input.

**Key insight**: Only ~50% acceptance rate. Must demonstrate REAL impact beyond just polluting prototype.

## Process

### Step 1: Find Object Manipulation Sinks

```
grep -rn "Object\.assign\|Object\.defineProperty\|Object\.create" .
grep -rn "merge\|extend\|deepMerge\|deepExtend\|deepAssign\|mixin" .
grep -rn "clone\|deepClone\|cloneDeep\|deepCopy" .
grep -rn "set\|setPath\|setValue\|lodash\.set\|_.set" .
grep -rn "\[.*\]\s*=" . --include="*.js"  # Bracket notation assignment
```

### Step 2: Check for Recursive Property Setting

Look for patterns where object keys from user input are used as property paths:
```js
// VULNERABLE: recursive merge without key filtering
function merge(target, source) {
  for (const key in source) {
    if (typeof source[key] === 'object') {
      target[key] = merge(target[key] || {}, source[key]);
    } else {
      target[key] = source[key];
    }
  }
}
```

### Step 3: Check Key Filtering

```
grep -rn "__proto__\|constructor\|prototype" . | grep -i "filter\|block\|skip\|ignore\|reject"
grep -rn "Object\.create(null)" .  # Null prototype objects are safe
grep -rn "hasOwnProperty\|Object\.keys\|Object\.entries" .
```

### Step 4: Assess Impact

Prototype pollution alone is often not enough. Look for impact:
- **DoS**: Polluted property causes TypeError crash (toString, valueOf)
- **Property injection**: Polluted property affects security logic (isAdmin, role, auth)
- **Gadget chains**: Polluted property reaches dangerous sink (eval, template)
- **Method clobbering**: toString/valueOf overwritten causing crash

## Dangerous Keys

| Key | Effect | Impact |
|-----|--------|--------|
| `__proto__` | Sets properties on Object.prototype | All objects affected |
| `constructor.prototype` | Same effect via constructor chain | All objects affected |
| `constructor` | Overwrites constructor reference | Type confusion |
| `toString` | Overwrites string conversion | TypeError on string operations |
| `valueOf` | Overwrites value conversion | TypeError on comparisons |
| `hasOwnProperty` | Overwrites property check | Logic bypass |

## CVSS Guidance

- Proto pollution + RCE gadget chain: CRITICAL 9.8
- Proto pollution + auth bypass: HIGH 8.1
- Proto pollution + DoS (crash): HIGH 7.5
- Proto pollution with no demonstrated impact: MEDIUM 5.3 (often rejected)

## References

- [Sinks](references/sinks.md) -- Object manipulation sinks
- [False Positive Indicators](references/false-positive-indicators.md)
- [PoC Skeleton](references/poc-skeleton.md)

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [ByamB4](https://github.com/ByamB4)
- **Source:** [ByamB4/find-cve-agent](https://github.com/ByamB4/find-cve-agent)
- **License:** Apache-2.0

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-byamb4-find-cve-agent-prototype-pollution
- Seller: https://agentstack.voostack.com/s/byamb4
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.