Loading vulnerability…
CVE-2026-7299 is a stored cross-site scripting vulnerability in Appsmith’s SQL query editor autocomplete functionality. In the CodeMirror-based SQL editor, database object names such as table or column names are not properly sanitized before being rendered into the DOM via innerHTML. An authenticated user with Developer access can create or influence malicious database object names so that when autocomplete suggestions are displayed to other users interacting with the same datasource, attacker-controlled script executes in their browser sessions. The issue is persistent because the payload is embedded in database metadata and is re-presented to users through the autocomplete renderer.
Mallory correlates every CVE against your assets, your vendors, and active adversary campaigns. Know which vulnerabilities matter for you, not just which ones are loud.
What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
If you can’t patch tonight, do this now.
Patch, then assume compromise.
1 valid exploit after Mallory filtered fakes, detection scripts, and README-only repos.
This repository is a small standalone exploit project for CVE-2026-7299, a stored XSS vulnerability in Appsmith SQL autocomplete. It contains two files: a README with vulnerability background, exploit flow, usage examples, and references; and a Python script (exploit.py) that automates exploitation. The exploit targets Appsmith's unsafe rendering of SQL autocomplete suggestions, where database table/column names are eventually inserted into the DOM via innerHTML. The attack model is web-based and stored: an authenticated attacker with Developer access creates a PostgreSQL table whose name is an HTML/JavaScript payload. When another workspace member later opens the SQL editor and triggers autocomplete for that datasource, the malicious table name is rendered and executes in the victim's browser. The Python script is operational rather than a mere PoC because it automates login, session/XSRF handling, workspace setup, and payload generation. The visible code shows authenticated interaction with Appsmith API endpoints /api/v1/users/me and /api/v1/login using a requests.Session. It supports three payload modes: a default alert(document.domain) XSS payload, a callback-based cookie exfiltration payload built around fetch(...document.cookie), and a fully custom payload supplied on the command line. The script also accounts for PostgreSQL's 63-character identifier limit and tries shortened callback variants, including protocol-relative URLs, to keep the malicious table name valid. Based on the README and visible code, the exploit's purpose is to weaponize a stored XSS path through Appsmith's datasource metadata/autocomplete pipeline, not just detect the issue. Successful exploitation can yield browser-context JavaScript execution as another Appsmith user, enabling session theft, API actions as the victim, and possible privilege escalation if the victim has elevated workspace permissions. Repository structure is minimal and focused: README.md documents the vulnerability and attack chain in detail, while exploit.py is the main entry point and likely contains the full automation logic for authentication, datasource discovery/creation, and malicious SQL execution.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
Vendor-confirmed product mapping. Mallory continuously reconciles this list against your asset inventory.
3 sources tracked across advisories and community write-ups. News coverage will land here when it surfaces.
No news coverage yet. Advisories and community discussion only.
Query your assets running an affected version, and investigate the blast radius.
Every observed campaign linking this CVE to a named adversary.
Malware families riding this exploit, with evidence and IOCs.
YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Cross-references every affected SKU, including bundled OEM variants.
Community discussion across Reddit, Mastodon, and other social sources.