Skip to content
SkillModel-led3,672 tokens in use

hunt-cors

A hunting guide for CORS misconfiguration built from nineteen public bug bounty reports, requiring a browser-proven credentialed read before reporting High.

14,533 stars on the repository MIT Licenseelementalsouls/Claude-BugHunter · skills/hunt-cors
/plugin marketplace add elementalsouls/Claude-BugHunter /plugin install claude-bughunter@elementalsouls
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing hunt-cors costs

98tokens idle
3,672tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

shelldetectednetworkdetectedwrites filesdetected

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Excellent
  • Frontmatter12/15
  • Structure21/25
  • Examples12/15
  • Output15/15
  • Size15/15
  • Resources10/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent
  • Runs shell commands−5

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 71
  • Reaches the network−10

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 72

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Excellent
  • Development activity100
  • Maintenance84
  • Community68
  • Maturity28
  • Momentum40

hunt-cors is a skill for Claude covering CORS misconfiguration: origin-reflection with credentials, null-origin trust, subdomain-regex bypass classes, pre-flight gating bypass, and postMessage origin checks. It opens with the two browser rules that kill most false findings: a wildcard Access-Control-Allow-Origin cannot carry credentials, and Access-Control-Allow-Credentials means nothing unless the origin is also reflected and the body is readable by a credentialed cross-origin fetch.

The skill’s regex-bypass section matches each specific parsing flaw (missing dot separator, missing end-anchor, unescaped dot) to the one payload that exploits it, keeping a tester’s payload targeted to the server’s actual flaw. Every High-severity claim requires a real browser proof of concept, since curl does not enforce the same-origin policy a browser does. It is for assets covered by a program or engagement you are authorized to test.

What You Get

  • A two-rule screening pass naming exactly when ACAO and ACAC are and are not exploitable for credentialed cross-origin reads
  • A regex-bypass table mapping five distinct subdomain-trust parsing flaws to the one payload each one requires
  • Browser HTML proof-of-concept templates for reflect-any-origin reads, null-origin reads via a sandboxed iframe, and trusted-subdomain reads
  • A pre-flight (OPTIONS) gating test for whether an attacker origin can drive state-changing PUT/DELETE requests with custom headers
  • A chain table and validation discipline section tying CORS findings forward into CSRF, XSS and subdomain-takeover chains

Common Use Cases

  • Testing whether an API endpoint reflects an arbitrary Origin header with credentials enabled, exposing an authenticated response to any origin
  • Identifying which specific subdomain-trust regex flaw a server has and sending the one bypass payload that matches it
  • Building a real browser proof of concept to confirm a CORS misconfiguration is exploitable, since curl alone cannot prove a credentialed cross-origin read

What Makes It Unique

Most CORS guidance treats every reflected origin as a finding and every bypass payload as worth trying against every target. This skill names the specific browser rules that make most header-level findings non-exploitable, and its regex-bypass table assigns one correct payload per parsing flaw, giving a tester a direct match for the server’s exact bug. It also states outright that curl output proves nothing about browser-enforced CORS, so every High-severity claim needs a browser-rendered proof.

Under The Hood

The skill is a single SKILL.md of guidance with no bundled scripts. Following it, the agent sends HTTP requests with varied Origin headers to API endpoints to read Access-Control-* response headers, and writes small HTML proof-of-concept files that it opens in a real browser to confirm whether a cross-origin credentialed fetch returns a readable body. Use it only against assets you are authorized to test.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers