Loading vulnerability…
CVE-2026-0651 is a path traversal vulnerability affecting TP-Link Tapo C260 v1, D235 v1, and C520WS v2.6. The flaw exists in the embedded HTTP server’s handling of GET request paths. According to the provided content, the server normalizes the request path before fully decoding URL-encoded input, and if normalization fails it falls back to using the raw path. This logic allows crafted URL-encoded traversal sequences to bypass intended directory restrictions. As a result, an attacker can request files outside the intended web root. The provided information indicates that authenticated attackers may be able to disclose sensitive system files and credentials, while unauthenticated attackers may at minimum probe for and access non-sensitive static assets or infer the existence of filesystem paths/files on the device.
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 Python proof-of-concept exploit chain targeting TP-Link Tapo C260 IP cameras, claiming unauthenticated-to-root RCE by chaining three CVEs (CVE-2026-0651/0652/0653). Structure: (1) exploit.py implements the cloud-based RCE portion: it sends HTTPS POST requests to the TP-Link cloud API endpoint /v1/things/{device_id}/services-sync with mobile-app-like headers and an Authorization token. It first 'poisons' a device configuration value by abusing a passthrough multipleRequest containing method setLedStatus; crafted nested JSON keys write attacker-controlled data into the config path tp_manage/info/dev_name (arbitrary config write). It then 'triggers' execution by sending another passthrough request (testUsrDefAudio -> device_info.set_region_code) which, per comments/README, causes the device to read dev_name and pass it unsanitized into a popen()-backed command path, resulting in command injection and privilege escalation to root. Payload options include a netcat-based reverse shell (mkfifo + /bin/sh -i | nc), a curl callback, or an arbitrary command string. The script optionally restores dev_name after exploitation. (2) lfd.py is a separate local-network helper for CVE-2026-0651: it performs an HTTPS GET to https://{host}/stok={token}/%2e%2e%2f... to traverse out of /www and read arbitrary files (e.g., /etc/passwd), requiring a valid stok token. requirements.txt pins requests/urllib3. Overall purpose: demonstrate and automate the exploit chain (config poisoning + trigger) and provide a standalone file disclosure tool; TLS certificate verification is disabled in both scripts.
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.
10 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.