Skip to content
KitploitKITPLOIT
ToolsExploitsBlog
Log in
Submit
ToolsExploitsBlog
Submit

Hacking, PenTest, and Cybersecurity Tools for Your Security Arsenal!

Kitploit is a directory of hacking, cybersecurity, and pentesting tools. Discover the latest project updates to find vulnerabilities, analyze systems, automate testing, and strengthen your security.

FeedsContactPrivacy© 2026 Kitploit

Tool Directory

Categories

View all categories
Loading categories
CVE-2026-43621 — Read-only verification harness for CVE-2026-43621, an SMF profile authorization bypass. Tests parameter variants to detect cross-account access to profile settings without exploiting it. | Kitploit
Tools/GitHubGitHub/kharaldipendra/cve-2026-43621
Authentication & AuthorizationDefensive ToolsVulnerability ScannersWeb Vulnerability ScannersVulnerability AnalysisWeb SecurityPenetration Testing
GitHub
Share
kharaldipendra/cve-2026-43621

CVE-2026-43621

Read-only verification harness for CVE-2026-43621, an SMF profile authorization bypass. Tests parameter variants to detect cross-account access to profile settings without exploiting it.

View Repository
23 days agoNot yet reviewed

CVE-2026-43621 — SMF profile authorization bypass

A read-only verification harness for CVE-2026-43621, an authorization flaw in Simple Machines Forum that lets a low-privilege member reach another member's profile account settings area.

The tool proves or disproves the missing authorization check. It does not contain an exploit chain, a password write, or an account-takeover step.


Authorized use

Run this only against a forum you own, or one you have written permission to test. The harness refuses any non-loopback host unless you pass --i-own-this-target, and it only sends GET requests plus an optional credential-free POST.


Vulnerability

CVECVE-2026-43621
ProductSimple Machines Forum (SMF)
AffectedSMF 3.0 — see Validation status
Fixed incommit 6f0dc61958aa (Sources/Actions/Profile/*)
ClassAuthorization bypass through a user-controlled key
ImpactCross-account access to profile settings, including the password-change form

The profile controller selects the member whose account form to render from a request parameter (u). The 3.0 refactor (Sources/Actions/Profile/*, with Profile::load()) does not bind that id to the authenticated session before the account-settings form is built. A member who is allowed to edit their own account can therefore ask for someone else's.

Because the returned form includes the new-password fields, reaching it is equivalent to account takeover — the target's password can be replaced through the normal save action. That write is intentionally not implemented here.

SMF 2.1.x uses a different profile implementation (Sources/Profile.php, no Profile::load). It is not affected by this code path. See Validation status.


What the harness does

  1. Control. Requests your own account area (u=<you>) and confirms it returns your own password form. If this fails, the run is inconclusive — the detector is wrong for the install, not the target safe.
  2. Matrix. Requests the target member's account area through a set of parameter variants that have historically confused PHP request parsing: duplicate keys in both orders, array syntax, mixed ;/& separators, null bytes, numeric coercion (01, 1.0, 1e0), and trailing whitespace.
  3. Verdict. A hit is a password form whose action targets the requested member id, returned while authenticated as a different member.

Every request is paced (--delay) and retried on 429/5xx.


Requirements

  • Python 3.8+
  • requests
pip install -r requirements.txt

Usage

python smf_profile_authz.py \
  --base http://127.0.0.1:8080/forums \
  --cookie-file .cookie \
  --my-id 42 \
  --target-id 1

--cookie-file holds the Name=Value pair copied from a logged-in browser session, for example:

SMFCookie240327=%7B%220%22%3A42%2C...

You can also pass --cookie "SMFCookie240327=..." directly, or set SMF_PHPSESSID if the install requires a separate PHP session cookie.

Options

FlagMeaning
--baseForum base URL (required)
--cookie / --cookie-fileSession cookie for the low-privilege account
--cookie-nameCookie name (default SMFCookie240327)
--my-idYour member id (required)
--target-idMember id to reach (required, must differ from --my-id)
--include-postAlso try the id in a POST body (no save, no credential fields)
--delaySeconds between requests (default 0.5)
--retriesRetries on 429/5xx (default 2)
--jsonEmit a machine-readable report
--i-own-this-targetRequired for any host that is not loopback
--insecureSkip TLS verification

Exit codes

CodeMeaning
0Not vulnerable
1Vulnerable
2Inconclusive, or usage/connection error

Reading the output

### control: own account area (u=2) ###
  [--] own account (control)      200 len=674     own-account

### matrix: target account area (u=1) ###
  [!!] canonical                  200 len=680     target-account
  [!!] array u[]                  200 len=680     target-account
  [--] null byte                  200 len=148     denied  errors=['You are not allowed to access this section.']
  ...

### result ###
  [!!] canonical: action=profile;area=account;u=1
  [!!] array u[]: action=profile;area=account;u[]=1

  VULNERABLE: 2 request(s) returned the account form for u=1 while authenticated as u=2.
  This proves the authorization check is missing; it is not an account takeover.

own-account on a target variant means the server ignored the parameter and rendered your own form — not a finding. Only target-account counts.

A clean matrix is not proof of a patch. Confirm the control hit, then diff the install against the fixed release.


Validation status

EnvironmentResult
Local mock, vulnerable routingVULNERABLE (exit 1)
Local mock, patched routingNOT VULNERABLE (exit 0)
Live SMF 2.1.4 installNOT VULNERABLE — expected: 2.1.x has no Profile::load / Sources/Actions/Profile
Live SMF 3.0 build (vulnerable)not tested

The detector itself is exercised end to end against both mock modes, so the control/matrix/verdict logic is sound. It has not been run against a live vulnerable 3.0 install. Treat it as a detector whose positive result needs manual confirmation, not as a validated exploit.

Limitations

  • Theme or profile-mod markup can change the form shape; the control request is there to catch that.
  • A POST-only variant of the bug is not covered unless you pass --include-post.
  • The matrix is a hypothesis set, not an exhaustive parser fuzz.
  • Custom arg_separator.input settings are covered only for ; and &.

Why there is no exploit chain

The step that turns a missing check into a takeover is a single POST with two password fields. Shipping that publicly would convert a vulnerability advisory into a one-command account-takeover tool for every unpatched 3.0 forum. The authorization gap is fully demonstrated by returning the form; the write adds no diagnostic value to this repo.

If you need to prove full impact — for a vendor report or your own lab — do the save manually against an instance you control, and document it in your write-up rather than in a public tool.


Mitigation

  • Upgrade to the release containing commit 6f0dc61958aa.
  • If you cannot upgrade immediately, restrict action=profile requests to admins when the u parameter differs from the session member id, at the reverse proxy or WAF. This is a stopgap, not a fix; the matrix above shows how many spellings of the parameter have to be denied.

References

  • NVD: https://nvd.nist.gov/vuln/detail/CVE-2026-43621
  • Fix: https://github.com/SimpleMachines/SMF/commit/6f0dc61958aa
  • SMF repository: https://github.com/SimpleMachines/SMF

License

MIT — see LICENSE.

Download Tool

Most Popular

View all →

Discover the most used tools by our community.

  • 1

    OsintHunter

    letox4/osinthunter

    🎯 The ultimate OSINT toolkit: investigation graph, confidence scoring, alias families, map, timeline and reports. 49 modules, 400+ sites, no API keys required.

    891
  • 2

    ghostlock-app

    yukonga/ghostlock-app

    GhostLock One-Tap Execution App (CVE-2026-43499)

    547
  • 3

    ARTEX

    baarai/artex

    LLM multi-agent autonomous penetration testing platform with planner/worker agents, asset and exploration graphs, MITM traffic recording, and vulnerability retesting.

    369
  • 4

    redforge

    cybersecurity91/redforge

    Browser-side red team toolkit with 102 offline tools for recon, payload generation, credential attacks, crypto, encoding and engagement reporting.

    366
  • 5

    looking-glass

    anonghosty/looking-glass

    CSIRT platform that ingests Shadowserver reports into MongoDB and replays them as interactive maps, analysis, findings, advisories and executive dashboards.

    309
  • 6

    CyberStrikeAI

    ed1s0nz/cyberstrikeai

    The system of action for AI-native cybersecurity—where intent becomes governed execution, evidence becomes operational memory, and every operation improves the next.

    282

Explore all tools

Browse our collection of tools

View all tools →