Skip to content
OpenAgentSkill
Browsing is open · Catalog captured 2026-11-21 · Automatic updates and account services are paused.
Security

access-control

Broken access control hub — horizontal/vertical privilege escalation, broken function-level authorization (BFLA), parameter-based and Referer/Location-based controls, multi-step process flaws, URL-matching/platform bypasses, and multi-tenant isolation. Tier 1. Builds a role×endpoint matrix. Use to systematically test authorization across an app/API.

AnandJogawade · Static checks on file

Installation

Read the source and dependencies before running a command. Copying does not install or execute anything.

npx skills add AnandJogawade/Zazy-Pentest-Plugin-for-Claude --skill access-control

About this skill


name: access-control description: Broken access control hub — horizontal/vertical privilege escalation, broken function-level authorization (BFLA), parameter-based and Referer/Location-based controls, multi-step process flaws, URL-matching/platform bypasses, and multi-tenant isolation. Tier 1. Builds a role×endpoint matrix. Use to systematically test authorization across an app/API.

Broken Access Control (hub)

Access control must be tested by trying to cross every boundary with at least one account per role. Build a role×endpoint matrix; every cell where actual access > intended is a finding.

Trigger phrases

access control, privilege escalation, bfla, function level auth, forced browsing, authorization bypass, multi-tenant

Vectors

  • Vertical (BFLA): replay admin/privileged requests as a low-priv/unauth user; forced-browse /admin, /api/internal/*, actuators, debug routes; guess sibling routes and verbs.
  • Horizontal: act on peer users' resources (→ idor).
  • Parameter-based: roles/flags in request (?admin=true, role=admin, JWT/cookie claim, hidden field) — tamper them (→ mass-assignment).
  • Referer/Location-based controls: app only checks Referer or hides UI — call the endpoint directly.
  • Multi-step flaws: skip or reorder steps; reach a later step without passing an earlier gate.
  • URL-matching / platform bypasses [NEW/RARE]: /admin/..;/, /%2e/admin, /admin%2f, trailing / or ., case, double-encoding, X-Original-URL/X-Rewrite-URL/X-Forwarded-For: 127.0.0.1 header overrides, Spring/Node/nginx-specific path normalization gaps.

Method

For each endpoint record intended-vs-actual access per role (unauth, user A, user B, admin). Test method tampering (GET↔POST, X-HTTP-Method-Override), content-type switch, and path-normalization bypasses on anything "protected."

Escalation

Vertical escalation → admin functions; chain with mass-assignment/idor for takeover. Note systemic (framework-wide) root causes.

Backend

Use the preferred Kali path (native Kali / WSL+Kali / Kali-SSH / Kali-MCP / HexStrike / Metasploit-MCP) or a headless browser for client-side steps; Windows-local only as last resort. See zazy-orchestrator.

Verification (background, before reporting)

Re-confirm every candidate independently (second account/session, replay, or OAST) before marking CONFIRMED. Report true positives only; hold back anything not reproduced.

Evidence

The role×endpoint matrix and two-context proof for each cross-boundary success.