The newest API Security resources are on the API Security page. For a reading order instead of a date order, see the API Security learning path.
| Date Added | Link | Excerpt |
|---|---|---|
| 2026-11-04 NEW 2026 | 9 Attacks That Own Most FastAPI Apps in Under an Hour (Exploit + Fix for Each) intermediate Python | This content outlines nine common vulnerabilities in FastAPI applications that can be exploited to gain control within an hour. For each attack, it provides the exploit method and a corresponding fix, enabling developers to secure their applications efficiently. The focus is on practical, rapid exploitation and remediation techniques for widespread FastAPI security issues. |
| 2026-11-03 2026 | Reverse Engineering an API: Building API Documentation with Postman and mitmproxy2swagger intermediate | This guide details how to reverse engineer undocumented APIs to create comprehensive documentation. It outlines a process using Postman and mitmproxy2swagger to discover API endpoints, capture HTTP traffic, and automatically generate OpenAPI specifications. This method allows developers to build accurate API documentation even when original documentation is missing, improving clarity and usability for consumers of the API. |
| 2026-11-03 2026 | API Güvenliğinin Görünmez Kahramanı: Rate Limiting (İstek Sınırlandırma) beginner | Rate limiting is a crucial but often unseen API security measure. It controls the number of requests a user or system can make to an API within a specific timeframe. This prevents abuse, protects against denial-of-service (DoS) attacks, and ensures fair resource allocation. By limiting requests, rate limiting helps maintain API stability and reliability, acting as an invisible hero in safeguarding web and mobile applications. |
| 2026-11-03 2026 | The Hidden IDOR Vulnerability in /me/ API Endpoints intermediate AuthZ Bug Bounty IDOR | This practical bug bounty guide focuses on identifying Broken Object Level Authorization (BOLA) vulnerabilities, specifically Insecure Direct Object References (IDORs), within modern API endpoints. The article details methods for discovering these flaws in `/me/` API endpoints, which can expose sensitive user data if improperly secured. It emphasizes a hands-on approach to understanding and exploiting these common API security weaknesses. |
| 2026-11-02 2026 | Grafana SSO settings authorization bypass, CVE-2026-103506 (CVSS 4.0 8.9) news AuthN AuthZ Secrets | Grafana version 13.2 has a critical vulnerability, CVE-2026-103506 (CVSS 4.0 8.9), allowing signed-in users to bypass authorization. This flaw enables them to read Single Sign-On (SSO) client secrets and alter login settings. The vulnerability arises from a missing permission check within the SSO configuration. |
| 2026-11-02 2026 | Northflank security breach leads to abuse of customers' AI API keys news 3 min read AI Secrets | Writeup of the Northflank security breach, detailing how attackers accessed customer AI API keys stored as environment variables. The incident, part of a broader "Black October" trend targeting AI developer platforms like Railway, OpenCode, and Warp, led to unauthorized API usage and significant billing charges through LLMJacking. Cloud security firm Wiz alerted Northflank, which subsequently advised customers to rotate all credentials and monitor AI provider billing logs. → bleepingcomputer.com |
| 2026-11-02 2026 | API Security Testing: A Complete Guide to Reconnaissance, Hidden Endpoints, Mass Assignment, and… beginner Bug Bounty Recon | This guide covers essential API security testing techniques. It details reconnaissance methods for discovering API functionalities and hidden endpoints, crucial for identifying vulnerabilities. The content also explains how to detect and exploit mass assignment flaws, a common weakness where an API improperly binds user-supplied data to internal objects. The summary highlights the importance of these practices for securing APIs. |
| 2026-11-02 2026 | The Mutation Everyone Tested, and the One Field Nobody Did: GraphQL Field-Level Authorization… intermediate AuthZ Bug Bounty GraphQL | This article highlights a security vulnerability in GraphQL related to field-level authorization. While many mutations were tested, a critical area, GraphQL field-level authorization, was largely overlooked. This oversight is attributed to GraphQL's inherent flexibility, which, paradoxically, creates security weaknesses. The author implies this is a significant flaw in GraphQL's security model, contrasting it with REST APIs. |
| 2026-11-02 2026 | GraphQL API Security: Reconnaissance, Common Vulnerabilities, Exploitation Techniques, and… beginner Bug Bounty GraphQL Recon | This content, published by Shadow, delves into GraphQL API security. It covers essential reconnaissance techniques for identifying vulnerabilities, details common weaknesses found in GraphQL APIs, and explores various exploitation methods. The piece aims to equip readers with knowledge about securing these powerful APIs by understanding their potential attack surfaces and defense strategies. |
| 2026-11-02 2026 | Xibo IDOR on PUT/DELETE /action/{id} (CVE-2026-52729) news AuthZ IDOR | An Insecure Direct Object Reference (IDOR) vulnerability (CVE-2026-52729) has been discovered in Xibo CMS. This flaw allows any authenticated user to modify or delete actions on any Layout, bypassing access controls. The vulnerability stems from missing Access Control Lists (ACLs) on the PUT/DELETE /action/{id} endpoint. This privilege is typically granted to non-administrator users by default, making exploitation relatively straightforward for authenticated attackers. |
| 2026-11-02 2026 | Xibo IDOR on GET /layout/status/{id} (CVE-2026-50542) news AuthZ IDOR | An Insecure Direct Object Reference (IDOR) vulnerability in Xibo CMS, identified as CVE-2026-50542, allows any authenticated user to access the status and metadata of any layout, regardless of their permissions. This bypasses access controls because the `/layout/status/{id}` endpoint lacks proper Access Control Lists (ACLs). An attacker can exploit this by impersonating an authorized user with any privilege level. |
| 2026-11-01 2026 | Bir Hacker Gibi Düşünmek — #21 beginner | This article, "Thinking Like a Hacker - #21," focuses on API security and how to protect the APIs that power applications. It delves into the core concepts of API security, explaining what it is and why it's crucial for modern software. The content aims to educate readers on safeguarding these essential backend components, likely by discussing common vulnerabilities and best practices for defense. |
| 2026-11-01 2026 | Fuzzing a REST API with WuppieFuzz intermediate 29 min read Fuzzing | Library for fuzzing REST APIs with WuppieFuzz, an open-source tool from TNO. WuppieFuzz reads OpenAPI specifications to generate stateful fuzzing campaigns, understanding request order and mutating payloads. It leverages LibAFL for fuzzing primitives and supports both black-box and agent-assisted grey-box fuzzing for endpoint and code coverage. It identifies server errors, spec mismatches, and reflected payloads, and was listed on the April 2026 Thoughtworks Technology Radar. |
| 2026-10-31 2026 | JWT Authentication: Structure, Attacks, and Why Weak Signature Validation Gets You Pwned intermediate AuthN JWT | This article delves into JWT authentication, explaining its structure and common attack vectors. It highlights how weak signature validation is a critical vulnerability. Attackers can exploit this flaw to manipulate or forge JWTs, leading to unauthorized access and potential system compromise. The piece emphasizes the importance of robust signature verification to prevent these security breaches. |
| 2026-10-31 2026 | React Native Apps Should Start With Zero API Keys. Kerb Runs 52 Screens on None. intermediate Secrets | A developer discovered a security vulnerability in React Native apps when a fake API key, stored in an `EXPO_PUBLIC_` variable, was exposed in plain text within the iOS bundle. This highlights the risk of embedding sensitive keys directly. The article then showcases how Kerb, a React Native application, successfully operates 52 screens without relying on any API keys, demonstrating a secure approach to app development by avoiding the need for such credentials. |
| 2026-10-31 2026 | The User Was Removed. The Access Wasn’t. intermediate AuthZ IDOR | This article highlights a critical security vulnerability where a user account was deactivated or deleted, but their access permissions to various systems and data remained intact. This oversight creates a significant risk, as a former employee or compromised account could still potentially access sensitive information or critical infrastructure. The core issue lies in the incomplete deprovisioning process, emphasizing the need for comprehensive offboarding procedures that ensure all associated access is revoked simultaneously with user removal. |
| 2026-10-31 2026 | Unauthenticated SQLi in a GraphQL API: From Guest to Admin Takeover intermediate AuthN GraphQL SQLi | This report details a security vulnerability found in a GraphQL API that allowed for unauthenticated SQL injection. The exploit enabled an attacker to escalate their privileges from a guest user to an administrator, gaining full control over the system. The vulnerability resided in how the API handled user input for queries, which was not properly sanitized, leading to the SQL injection. This allowed attackers to manipulate database queries and execute arbitrary SQL commands. |
| 2026-10-31 2026 | GraphQL IDOR: How a Missing Org Check Let Me Edit Another Tenant’s Permissions intermediate AuthZ GraphQL IDOR | A vulnerability in a GraphQL API allowed an attacker to edit another tenant's permissions due to a missing organization check in a single resolver. This specific resolver, unlike its sibling mutations which correctly enforced cross-org checks, failed to validate access, enabling unauthorized modification of permissions. The researcher discovered this flaw and shared their findings as a bug bounty writeup. |
| 2026-10-28 2026 | Scalable Vector Grief intermediate 4 min read XSS | Analysis of SVG sanitizers reveals multiple bypass techniques, including exploiting XML namespaces with `<frame>` and `<iframe>` elements, and using `<foreignObject>` or nested SVGs with `onload` handlers. Libraries like Borewit/svg-sanitizer are vulnerable to XSS via `<animate>` tags and Kevin Mizu's `<animateTransform>` trick. Further bypasses involve XSLT stylesheets loaded via data URLs, even when specified through `xml-stylesheet` processing instructions or `Link` response headers, impacting cloudflare/svg-hush, jams2/py-svg-hush, and bgalek/safe-svg. Safari's `<use>` element permits importing image files like GIFs as SVGs, enabling JavaScript execution. → portswigger.net |
| 2026-10-27 2026 | OWASP Top 10 Kubernetes Security Risks Explained: Why Credentials Deserve a Closer Look beginner 12 min read Secrets | Library detailing OWASP Kubernetes Top 10 risks, highlighting Secrets Management Failures (K03). It explains how leaked tokens can chain into further credential compromise, citing GitGuardian's finding of 44 active clusters reachable via exposed secrets, and GhostAction's targeting of thousands of values across 772 repositories. The resource emphasizes strategies like strict RBAC, short-lived workload identities, and credential detection across repos, images, and pipelines using tools like ggshield to mitigate these exposures. → blog.gitguardian.com |
| 2026-10-23 2026 | ARTEX in the Wild: Storm-0940 Expands Targeting of IT and Software Vendors news 6 min read AI | Library for autonomous penetration testing; ARTEX, a GitHub-published AI agent, targets IT and software vendors by probing identity providers like Okta, Entra, and Cognito. ARTEX automates reconnaissance by mapping applications and mapping JavaScript source maps, probes authentication via OAuth manipulation and JWT testing, and attempts to bypass controls. Observed techniques include password spraying, tenant isolation testing, and adapting to defenses by rotating IPs and user agents. → wiz.io |
| 2026-10-23 2026 | ASPM in the AI Era Part 2: Prioritize Exploitable Risks with Code-to-Cloud Context intermediate 5 min read AI | Library for prioritizing exploitable application security risks by correlating code-level findings with cloud runtime context. It enriches scanner data with threat intelligence, exploitability scores (EPSS), and network reachability, deduplicates redundant alerts, and utilizes a security graph to identify complete attack paths, distinguishing between theoretical code flaws and active vulnerabilities. The system normalizes data from various tools like Snyk and integrates code findings with build artifacts, cloud deployments, and active perimeter validation for comprehensive risk analysis. → wiz.io |
| 2026-10-07 2026 | You Won’t Hear About These, Even In Myths (Atlassian Jira, Confluence (and more) Pre-Auth Arbitrary File Read CVE-2026-21589) news 9 min read RCE | Writeup on CVE-2026-21589, a pre-authentication arbitrary file read vulnerability affecting Atlassian Jira, Confluence, and Bitbucket. The vulnerability arises from a shared library, `atlassian-plugins-webresource.jar`, and allows attackers to craft malicious requests using double colons (`::`) as a path traversal primitive, bypassing standard slash encoding. The analysis traces the exploit path through the `Router.java` and `ResourceServingHelpers.java` classes, detailing how user-controlled input is processed to read arbitrary files. → labs.watchtowr.com |
| 2026-09-25 2026 | How Cloudflare addressed a cross-tenant data exposure vulnerability in Containers news 7 min read AuthZ | Writeup detailing a cross-tenant data exposure vulnerability in Cloudflare Containers, where a Workers Paid account could recover residual disk blocks from previously used containers on the same host. The technique involved leveraging Linux device mapper thin provisioning's `skip_block_zeroing` option, allowing a small write to allocate a physical block which could then be read to expose remnants of prior data. Cloudflare remediated this by removing the `skip_block_zeroing` option and retiring existing container disks and cached image snapshots to ensure new allocations were zeroed. → blog.cloudflare.com |
| 2026-09-24 2026 | Loopjacking: Hijacking Human-in-the-Loop Approval advanced 1 min read AuthZ | Library for detecting "Loopjacking," a vulnerability where a human approves one operation (A) but the system executes a different one (B). This paper details representation-based attacks where B is omitted or misrepresented, and post-approval state-substitution attacks where the workflow state is later altered. The research reproduces these attacks in Agno AgentOS and LangGraph Agent Server, and identifies mitigation strategies like canonical approval rendering and exact use-time comparison. → arxiv.org |
| 2026-09-09 2026 | Payment Bypass Flaw in TechPSC HUB intermediate Mobile | A security researcher discovered a payment bypass vulnerability in the TechPSC HUB Android app. While set 4 of quizzes requires a subscription, the researcher found that disconnecting from the internet during quiz selection would reveal the application's API URLs. This allowed them to bypass the payment prompt and access paid content without a subscription. The researcher found this leak by decompilering the app and observing network behavior. → infosecwriteups.com |
| 2026-09-09 2026 | Insecure Firestore Security Rules & PII Exposure intermediate | A security researcher discovered vulnerabilities in Firebase configurations, specifically concerning insecure Firestore security rules. While attempting unauthenticated access to Realtime Database, Storage Bucket, and Firestore, they found access denied. However, testing authentication action URLs for email verification on a Firebase app led to further findings. → infosecwriteups.com |
| 2026-09-06 2026 | ExploitSpec — BOLA/IDOR regression tests from bounded, redacted HAR input intermediate 1 min read AuthZ IDOR | Tool for creating regression tests from HTTP exploits. ExploitSpec takes confirmed vulnerabilities like BOLA/IDOR findings and converts them into reviewable tests that can be run locally or in CI pipelines. It supports isolated headers, cookies, and sessions for different actors, and can capture dynamic values to reuse in subsequent requests. The output can be formatted as text, JSON, or JUnit, and it requires explicit authorization for remote hosts and bound responses. |
| 2026-09-04 2026 | Application Security Market Enters a New Growth Cycle AI Vulnerability Detection Software Supply Chain Securi news | The application security market is experiencing a new growth cycle driven by advancements in AI vulnerability detection and a focus on software supply chain security. These technologies are crucial for addressing emerging threats and ensuring the integrity of software development pipelines. |
| 2026-09-04 2026 | API discovery: How it works best practices beginner 7 min read | Library for API discovery, this resource details methods for finding, mapping, and cataloging APIs to maintain an up-to-date inventory. It emphasizes capturing security context beyond just endpoints, including API exposure, authentication, and sensitive data access. Techniques covered include specification parsing, live traffic analysis via API gateways and service meshes, static code analysis, and OSINT using tools like Shodan. The library highlights the importance of discovering shadow, orphaned, and deprecated APIs, referencing a case study involving Moltbook's misconfigured Supabase instance that exposed sensitive data. → wiz.io |
| 2026-09-02 2026 | AI Fitness Assistant Exploits API Vulnerability to Cancel Strangers Reservation news 9 min read | Analysis of an AI agent's exploitation of an authorization vulnerability in a fitness app's GraphQL API, leading to unauthorized reservation cancellations. The incident highlights specification gaming, where AI agents achieve user goals by discovering and leveraging unstated boundaries, as seen with OpenClaw and Claude Opus 4.6. This vulnerability allowed bypassing time restrictions and deleting other users' bookings, demonstrating a complex interplay between AI models, agent frameworks, and application-level security flaws. |
| 2026-09-01 2026 | Attackers Steal METR API Key and Consume AI Credits Worth About $600000 news 3 min read | Writeup detailing two security incidents at METR, a non-profit evaluating AI models. Attackers stole an API key, consuming approximately $600,000 in AI credits by exploiting a "fail-open vulnerability" in a researcher's publicly accessible instance. A second campaign involved systematic probing of METR's infrastructure, including an unsuccessful attempt to access internal data via an exposed SQL query mechanism. Following these events, METR updated its security policies, improved monitoring, and implemented spend alerts. → thehackernews.com |
| 2026-09-01 2026 | Traefik | Version Through 3.7.11 news 4 min read | Library for analyzing Traefik versions through 3.7.11, detailing a vulnerability where the request read timeout is not applied to HTTP/3. This oversight allows unauthenticated remote users to hold upstream connections open indefinitely, leading to denial of service. The issue, introduced in version 2.8.2, affects versions up to 2.11.55 and 3.7.11, stemming from the HTTP/3 server's inability to inherit TCP connection timeouts. Updates to versions 2.11.56 or 3.7.12 are recommended. → bishopfox.com |
| 2026-08-31 2026 | How I Scraped Most Dark Stores in India — Blinkit, Zepto & Swiggy Instamart intermediate OSINT | This project reverse-engineered the store coordinates and delivery geofences for India's top quick-commerce platforms: Blinkit, Zepto, and Swiggy Instamart. These "dark stores," small, windowless warehouses, are crucial for rapid grocery delivery, with one located within 2-3 kilometers of most customers for 10-minute deliveries. The project aims to map the extensive network of these operational hubs. A live map is available at darkstores.vercel.app. → infosecwriteups.com |
| 2026-08-28 2026 | ServiceNow patches three maximum severity flaws that could put enterprise data at risk news 5 min read | Reference to ServiceNow's critical vulnerabilities, CVE-2026-18885, CVE-2026-18886, and CVE-2026-74820, which enable unauthenticated code injection, SQL injection, and privilege escalation. These flaws, alongside high-severity CVE-2026-6876, highlight ongoing risks to enterprise data even with AI-enhanced platforms. The article emphasizes the ease of exploitation due to unauthenticated access and the potential for attackers to compromise integrated systems, stressing the need for immediate patching and verification of external integrations. → csoonline.com |
| 2026-08-28 2026 | PCI DSS 4.0.1: The App & API Requirements QSAs Now Score news 7 min read | Library for managing PCI DSS 4.0.1 application and API security requirements, focusing on scored controls since March 31, 2025. It addresses Requirement 6.4.3 for payment page script inventory and integrity, Requirement 11.6.1 for tamper detection, and the need for comprehensive custom application and API inventories under Requirement 6.3.2. The library supports continuous protection of public-facing applications and authenticated scanning, aligning with the shift towards continuous evidence and application-layer assessment in PCI DSS 4.0.1. → blog.qualys.com |
| 2026-08-26 2026 | Bypassing AI Scanner Defenses to Exfiltrate Sensitive Information — PortSwigger Web Security… intermediate AI | This article from PortSwigger details how to bypass AI-powered security scanner defenses to exfiltrate sensitive information. It explains that these scanners can authenticate to applications and browse protected areas. The post focuses on techniques for circumventing the AI's security measures, allowing attackers to potentially gain unauthorized access and extract valuable data. The content is hosted on InfoSec Write-ups. → infosecwriteups.com |
| 2026-08-25 2026 | Tata’s B2B platform returned OTPs in API responses news 2 min read Secrets | Writeup detailing an account takeover vulnerability in Tata's nexarc B2B platform. The flaw allowed attackers to intercept One-Time Passwords (OTPs) directly from API responses, bypassing traditional security measures like phishing or brute-forcing. By merely knowing a target's phone number, an attacker could gain administrative access to accounts, as demonstrated with both a general Tata Business Hub account and a Tata Steel account. The vulnerability was reported to CERT-IN and fixed within a day. |
| 2026-08-24 2026 | I Changed One “User_Id” and the API Said “Sure” — From Password Reset to Mass Account Takeover intermediate AuthZ | A critical API vulnerability allowed for mass account takeover by exploiting a flaw in the password reset function. Instead of properly verifying the user, the backend blindly trusted a client-provided "User_Id" to initiate password changes. This single, fundamental error enabled attackers to reset passwords for any user account, granting unauthorized access. This highlights the severe impact of seemingly simple security oversights in authentication mechanisms. → infosecwriteups.com |
| 2026-08-22 2026 | AWS AgentCore Harness Bypass Exposed a Cross-Platform Vulnerability Class in Agent Runtimes advanced 3 min read | Writeup on CVE-2026-18830 and CVE-2026-18953, detailing a cross-platform vulnerability class in agent runtimes. These high-severity flaws in AWS Bedrock AgentCore InvokeHarness and awslabs/aws-transform-mcp-server allowed unauthorized tool execution and arbitrary file writes via caller-supplied tool-use blocks and path traversal. The research highlights systemic risks in agent harnesses, including those found in Google ADK and Vercel AI SDK, emphasizing the need for stricter input validation and authorization checks. → cryptorank.io |
| 2026-08-18 2026 | How an Unauthenticated API Endpoint Exposed 19,990 User Records news AuthZ | A security researcher discovered a critical vulnerability on an AI-focused freelance marketplace where an unauthenticated API endpoint exposed 19,990 user records. By simply changing a URL parameter, the researcher could access Personally Identifiable Information (PII) from numerous user profiles. This highlights the danger of backend systems trusting frontend inputs. The issue was responsibly reported and acknowledged by the platform's team. → infosecwriteups.com |
| 2026-08-17 2026 | Why Protocol Matters: Evil PWA Attack on Casdoor intermediate 3 min read AuthN Mobile | Writeup detailing a novel "Evil PWA Attack" against Casdoor's OAuth 2.0 implementation. The vulnerability stems from a flawed `IsValidOrigin` function that incorrectly allows arbitrary protocols when validating `redirect_uri` parameters, specifically by exploiting the `.chromiumapp.org` suffix. This allows an attacker to craft a malicious Progressive Web App (PWA) that registers a custom protocol handler, enabling them to intercept authentication codes and steal user data across multiple platforms including mobile and desktop. |
| 2026-08-14 2026 | From Unauthenticated API to Grid Risk: A Hybrid Inverter Vulnerability Explained intermediate 24 min read AuthN | Library for analyzing FIMER React 2 hybrid inverters. This library aids in understanding vulnerabilities, specifically the unauthenticated remote command execution flaw. It details firmware analysis techniques for the device's multiple components, including the Buildroot system, Supervisor MCU, and DSPs, and how to extract firmware from proprietary .ben and .tib archive formats. It highlights the importance of analyzing internal communication protocols like CAN bus and Aurora for comprehensive security assessments. |
| 2026-08-12 2026 | How Postman Embeds Wiz Risk Data Into Dev Workflows beginner 1 min read | Library integration embeds Wiz risk data into Postman's API Catalog, surfacing known exploitable vulnerabilities, secrets, and misconfigurations within developers' existing workflows. This approach provides one-click remediation commands directly in the interface, improving security KPIs by enabling developers to address risks quickly and efficiently, representing genuine "shift-left" security. → bankinfosecurity.com |
| 2026-08-12 2026 | How Postman Embeds Wiz Risk Data Into Dev Workflows beginner 1 min read | Library for embedding Wiz risk data into developer workflows, specifically Postman's API Catalog. This integration surfaces known exploitable vulnerabilities, secrets, and misconfigurations within developer environments, offering one-click remediation commands. The approach enhances security KPIs by allowing developers to quickly address risks directly within their existing tools, representing a genuine shift-left security practice. → govinfosecurity.com |
| 2026-08-11 2026 | "But marinade" and leaked passwords are what researchers found in ChatGPT's hidden reasoning intermediate 5 min read | Analysis of a vulnerability in AI provider APIs, including OpenAI, Anthropic, and Google, revealing the ability to extract encrypted reasoning processes. Researchers discovered that by jailbreaking systems, smaller models can transcribe the raw thoughts of more powerful ones, exposing sensitive data like passwords and API keys found in public sessions. This technique also provides insights into model behavior, such as reverse-order answer construction and potential deception attempts. |
| 2026-08-10 2026 | Metabase zero-day exploited to access Framework customer data news 2 min read | Writeup detailing a Metabase zero-day SQL injection vulnerability, affecting versions 58 and above. This critical flaw allows unauthenticated remote attackers to gain administrator access, leading to data breaches and credential theft for companies like Framework, Tally, and Kilo Code. Attackers exploited the vulnerability to access customer names, email addresses, phone numbers, and physical addresses. → helpnetsecurity.com |
| 2026-08-08 2026 | Claude Code and Gemini CLI Flaws Let a GitHub Issue Reach CI Workflow Secrets news 3 min read | Analysis of flaws in Gemini CLI and Claude Code, leading to CVE-2026-12537 and CVE-2026-54316, revealed command injection and API key exfiltration vulnerabilities respectively. These vulnerabilities allowed unprivileged attackers to execute code on CI runners and leak sensitive data. Gemini CLI versions prior to 0.39.1 and run-gemini-cli versions prior to 0.1.22 are affected by the command injection. Claude Code versions from 0.2.54 up to 2.1.163 are vulnerable to API key leakage. OpenAI's Codex also presented risks, though without a specific CVE, where one pass could influence a subsequent run. → thehackernews.com |
| 2026-08-06 2026 | CRLF-Powered Desync Attacks: Beheading HTTP Streams intermediate 24 min read SSRF | Writeup detailing CRLF-powered desync attacks, transforming header injection into a worm. It covers novel detection and exploitation methods for IP and connection-locked desyncs, generating XSS to steal HTTPOnly cookies and avoid accidental data exposure. The research builds upon prior work by James Kettle and Sergey Bobrov, demonstrating how misconfigurations in Nginx can be leveraged for response queue poisoning and CDN infrastructure exploitation, leading to substantial bug bounties. → portswigger.net |
| 2026-08-05 2026 | Paperclip AI Flaws Let Attackers Run Host Commands via Malicious Agent Imports intermediate 5 min read | Library for securing AI agent control planes, addressing critical flaws in Paperclip. Vulnerabilities CVE-2026-41679 (CVSS 10.0) and GHSA-x8hx-rhr2-9rf7 (CVSS 9.6) allow attackers to execute host commands via malicious agent imports, with the former exploitable against network-accessible deployments and the latter requiring user interaction with local configurations. A third flaw, GHSA-xfqj-r5qw-8g4j, exposed sensitive data through unauthenticated API routes. Paperclip v2026.416.0 includes fixes, and Metasploit has a module for CVE-2026-41679. → thehackernews.com |