Loading vulnerability…
CVE-2026-5530 is a server-side request forgery vulnerability in Ollama model-pull processing affecting releases through 0.18.1. An unauthenticated remote requester can supply a manipulated model source that causes the Ollama service to make outbound HTTP requests to attacker-selected destinations. The affected non-tensor blob download processing can follow redirects to a different host, permitting requests to internal services or external systems not directly reachable by the attacker.
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.
Repository contains a single Python proof-of-concept exploit (poc.py) and a README describing usage. The exploit targets CVE-2026-5530 in Ollama by standing up two attacker-controlled HTTP services that emulate OCI/Docker registries. The first service ('Rogue') listens on a configurable port (default 9999) and serves minimal registry metadata, then issues an HTTP 307 redirect on blob requests to an attacker-selected target URL. This causes Ollama to perform SSRF to arbitrary internal HTTP endpoints reachable from the Ollama host. In enum mode, the script measures whether pulls succeed or return quickly to identify live internal endpoints. In probe mode, it binary-searches the expected Content-Length by varying the declared blob size and timing pull behavior, allowing the attacker to infer the size of the SSRF response. In exfil mode, once the response size is known or probed, the script abuses Ollama's /api/copy and /api/push endpoints to push the captured blob contents to a second attacker-controlled registry ('Exfil', default port 10000), where the response body is reconstructed and printed. The code is operational and weaponized enough to perform real SSRF enumeration and response exfiltration, but it is a standalone PoC rather than part of a larger exploitation framework.
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, community write-ups, and news. New activity surfaces here as Mallory finds it.
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.