CVE-2026-22874 is a server-side request forgery vulnerability in Gitea affecting versions up to and including 1.26.2. The flaw is caused by incomplete destination filtering in webhook handling and repository migration allow-list logic. The protection relied on Go standard library IP classification and did not comprehensively block internal and special-use address ranges, including at least RFC 6598 Carrier-Grade NAT space. As a result, an authenticated user who can configure or trigger outbound requests through webhooks, and potentially repository migration features, can cause the Gitea server to send HTTP requests to otherwise restricted internal services or cloud metadata endpoints. The issue is particularly severe because webhook delivery history can expose the HTTP response, including headers, status, and response body, making the SSRF non-blind and enabling direct retrieval of sensitive data such as cloud instance metadata credentials.
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 (1 hidden).
This repository is a small standalone Python PoC for CVE-2026-22874 affecting Gitea webhook handling. The repo contains one executable Python script and a README. The script defines a GiteaSSRFExploit class that authenticates to a target Gitea instance using a personal access token, creates a repository webhook with a user-controlled destination URL, then triggers webhook execution by modifying the repository. Two trigger paths are implemented: a git-based path that clones the repository, creates a file, commits, and pushes; and an API-based path that creates repository content directly through the Gitea API. After triggering, the exploit waits and polls webhook deliveries to determine whether the outbound request occurred, making it an operational SSRF validation/exploitation tool rather than mere detection. The CLI accepts the target URL, token, owner, repo, bypass profile, trigger method, and wait time. The visible code shows several named SSRF bypass profiles (nat64-imds, azure, cgnat, ipv4-mapped, teredo, 6to4), indicating the exploit is designed to test alternate address encodings or routing tricks against webhook destination filtering. No destructive behavior or post-exploitation payload is present; the main capability is inducing server-side requests to attacker-selected/internal endpoints via Gitea webhooks.
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.
8 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
An SSRF vulnerability in Gitea webhook handling caused by incomplete internal IP filtering, enabling access to cloud metadata services and potential theft of AWS, Azure, or GCP instance credentials.
An SSRF vulnerability in Gitea webhook handling caused by incomplete internal IP filtering, enabling access to internal services and potential cloud metadata credential theft, including AWS IMDS.
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.