Urgent.News

What's breaking now, across thousands of outlets.

Tech

HotGo Let the Client Choose the SMS Login Verification Code

HotGo Let the Client Choose the SMS Login Verification Code An SMS verification code is useful only when the server controls its value and ties it to one login attempt. The GitHub Advisory Database describes CVE-2026-120447 as a HotGo authentication bypass where the client can choose that code. The affected versions are HotGo 2.8.4 through 2.15.1. The advisory rates the issue Critical with a CVSS…

HotGo's authentication process is at risk due to a vulnerability allowing clients to select the SMS login verification code. This issue, described as a critical authentication bypass, affects versions 2.8.4 through 2.15.1 of HotGo. The advisory rates it with a CVSS score of 9.2.

An attacker can exploit this by providing a known code value through a server-controlled field. If they know the user's mobile number, they can complete a login as that user. This vulnerability is not related to SMS delivery; the server accepts the client's chosen code instead of generating a unique one. The trust boundary is broken before the verification code leaves the server.

To exploit this, an attacker requires network access to the affected endpoint and a target mobile number. The exploit does not require user interaction. High-risk areas include public and reseller administration portals, with internal management portals being relevant if an attacker has access to a user station or network. Signs of account takeover include unusual authentication logs, changes in user credentials, and missing audit logs.

To remediate this issue, administrators should upgrade to the hotfix referenced in the advisory. If upgrading isn't possible, disabling SMS login for privileged accounts, requiring a second authentication method, and restricting administration endpoints to trusted networks are advised. Rotating sessions and credentials for potentially compromised accounts is also recommended.

After the fix, it's crucial to verify that the server still ignores the client-supplied code and generates a proper, time-limited code tied to the correct mobile number. This verification is necessary as a patch might only address one login flow while leaving others vulnerable.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

TE 1888247 is an SFP+ connector, not an SFP cage: what to match when you need an alternative

TE Connectivity part number 1888247 is a 20-position SFP+ board-side connector (receptacle) , not an SFP cage. On TE's product pages, both 1888247-1 and 1888247-2 are described as "SFP+, Connector…

  • TE 1888247 is a 20-position SFP+ connector, not a cage
  • Two variants exist: 1888247-1 and 1888247-2, differing in gold plating thickness
  • Alternatives require matching 20-position SFP+ connector and compatible cage

Autonomous attacks with Cairn hit 10 organizations

Autonomous attacks with Cairn hit 10 organizations out of 19 marked as targets. Unit 42, the threat intelligence team at Palo Alto Networks, published the report on November 12: between August and…

  • Autonomous attacks using Cairn framework target 10 out of 19 organizations
  • Chinese-speaking attackers conduct 25 operations, accessing data from over half
  • Operations automated, lasting 20 minutes to over 50 hours, targeting defense and defense entities

Agent Access: How Do You Scope an Agent's Permissions Safely?

I used AI tools to help draft and edit this post. I checked the facts and the wording, and the original lives at geekindata.com . The diagrams are on that page. Front door. Catalog. Corpus.

  • Implement sign-in group for shell with limited access, not full permissions.
  • Backend app roles enforce agent access and knowledge category access on sensitive calls.
  • SPA consumes capability endpoints, not parsing roles from JWT, to maintain single access payload.

Your coding agent's "done" is a claim. Plan the checks between steps

Most agent workflows I see are not one request. They are a run of them: add the model, wire the endpoint, add validation, write the migration, build the report.

  • Agents without checks complete 51% of 20-step sequences
  • Checked agent code matches human reliability, but 14-33% failures remain
  • Separate agent checks needed, regression runs maintain code quality

More from Friday 13 November →