< ciso
brief />
Tag Banner

All news with #api security tag

51 articles

AI Agents Are Live in Production Before Security Is Ready

🔐 A Postman survey of nearly 9,000 developers, architects and executives shows that a quarter of organizations already have AI agents calling their APIs in production. Yet about four in ten have not worked out how identity and credentials should work for those agents. Three in ten let agents reuse the same credentials as ordinary API clients, leaving no separate identity to restrict or revoke. Unauthorized access to data and endpoints is the top concern.
read more →

Envoy Evolves into an Open Standard for Agentic AI Traffic

🤖 Google Cloud is positioning the open-source Envoy proxy as a foundation for AI traffic, MCP tooling, and agent-to-agent communication. New capabilities address the limits of traditional Layer 7 proxies, including payload blindness, request-based rate limiting, and stateless load balancing. Envoy now offers native body parsing, token-aware rate limiting, model and MCP routing, and inline security guardrails for autonomous agents.
read more →

Amazon API Gateway adds native JWT authorizers

🔐 Amazon API Gateway now supports a native JWT authorizer for REST APIs, enabling the gateway to validate JSON Web Tokens directly without requiring a Lambda authorizer. You can configure issuers, accepted audiences, allowed signing algorithms, and additional claims to validate, with the gateway checking signature, issuer, audience, expiry, and claims on each request. The authorizer can be attached to multiple methods with per-method scope requirements and supports Regional, edge-optimized, and private REST APIs across commercial AWS Regions and AWS GovCloud (US).
read more →

Trusted Agent Delegation and Web Intermediation

🛡️ This post explains how remotely hosted AI Agents act as intermediaries between Users and Sites, eroding Sites' ability to identify, serve, and hold users accountable. It describes the risks Agents introduce—traffic commingling, fraud, and loss of personalization—and argues Sites need new identity, reputation, and accountability tools. The post surveys approaches like Web Bot Auth, MOLE, and preference expression work, and calls for interoperable protocols to preserve user–site relationships.
read more →

AWS WAF adds Stripe fiat payouts for AI traffic

🔒 AWS WAF now integrates with Stripe so publishers, content owners, and API providers can collect stablecoin payments from AI bots and have proceeds settle in local currency directly to their Stripe balance. The feature removes the need for customers to hold stablecoins or manage crypto wallets, and Stripe appears as a payout option alongside existing stablecoin payouts. Connect via Stripe Connect in the AWS WAF console; existing pricing and agent policies remain unchanged.
read more →

Compute Engine API multi-versioning launched

🔔 Compute Engine API multi-versioning is now generally available, introducing date-based, immutable API snapshots alongside the existing /v1 endpoint to let teams adopt modern semantics without breaking production. Stable releases (YYYY-MM-DD) are quarterly and frozen, while preview releases (YYYY-MM-DD-preview) provide short-lived testing branches. Existing /v1 support remains uninterrupted, and clients can target versions via header or query parameter.
read more →

CloudWatch adds API key (bearer token) support for OTLP

🔐 Amazon CloudWatch now accepts bearer token authentication via API keys for sending OpenTelemetry traces to the AWS X-Ray OTLP traces endpoint. Any OTLP/HTTP client that can set an Authorization header — including the OpenTelemetry Collector and SDK exporters — can use a bearer token instead of AWS Signature Version 4 (SigV4). SigV4 remains the recommended default and ADOT SDKs continue to use it; both methods provide the same CloudWatch capabilities and bearer token support is available in all commercial regions where the endpoint exists.
read more →

Choosing a WAAP in 2026: Three decisive questions

🛡️ Buyers in 2026 face consolidated WAAP choices as platforms aim to replace separate WAF, API, bot, DDoS, and client-side controls with a single service. New AMTSO-certified test data from SecureIQLab v5.0 highlights large variance in overall protection and false-positive rates across vendors, making efficacy and tuning effort critical selection criteria. The article recommends probing API protocol coverage, time-to-tune on real traffic, and deployment/licensing scalability during proof-of-value trials.
read more →

Microsoft guidance to restore broken Exchange EWS apps

🛠️ Microsoft explains how tenant admins can re-enable the retired Exchange Web Services (EWS) API in Exchange Online to restore apps that stopped working after EWS disablement. The company details two remediation methods: turning off the Baseline Security Mode block for EWS in the Microsoft 365 admin center or running Set-OrganizationConfig -EWSEnabled $true in Exchange Online PowerShell, with changes taking effect within hours. Microsoft reiterated that EWS will be blocked by default on October 1, 2026, and fully disabled in April 2027, and advised app developers to migrate to the Microsoft Graph API as the long-term solution.
read more →

Networking architectures for AI inference model serving

🔒 This post compares two reference networking architectures for AI inference model serving: one tailored for Google Kubernetes Engine (GKE) and one for mixed or alternative backends. It explains a common control-plane pattern using Private Service Connect, optional Apigee, and Model Armor as a centralized entry point for secure, private inference calls. The GKE design adds a specialized GKE Inference Gateway, inference pools, and replica sets for GPU/TPU workloads. The multi-backend design uses a regional internal Application Load Balancer, a Cloud Run payload processor service extension to inject model headers, and Network Endpoint Groups to route to heterogeneous backends.
read more →

Cloudflare Application Profiles for Positive Security

🔒 Cloudflare is launching Application Profiles to enforce positive security by learning the expected structure and format of HTTP requests and flagging deviations. The feature extends existing API Schema Learning to web applications, creating per-operation profiles that validate paths, query parameters, headers, cookies, and request bodies. Profiles are learned from observed traffic, updated weekly, and produce metadata (cf.schema_validation.learned.violated) that teams can use in Security Rules to monitor or enforce blocking. A closed beta is available to invited Enterprise customers and those with API Security already have access.
read more →

Nearly 1 in 10 LiteLLM Gateways Exposed Default Key

🔒 Wiz Research found that many internet-facing LiteLLM gateways still accept the example admin key sk-1234 from the setup guide, allowing full admin access. The master key controls admin rights and authentication; if left default or unset, attackers can retrieve provider API keys and potentially cloud IAM credentials via pass-through endpoints. Several CVEs affecting guardrails, MCP authentication, and sandbox escapes have been fixed in recent releases, and upgrades plus key rotation are recommended.
read more →

Data Agent Kit: Agentic Analytics in IDEs

🧭 The Data Agent Kit streamlines exploratory data investigations by enabling agents inside IDEs to query multiple systems—data warehouses, operational databases, and object stores—via the open Model Context Protocol (MCP) and modular skills. It integrates with popular IDEs and plugins so developers can run queries, inspect execution trails, and grant scoped permissions without switching tools. The kit can convert ad-hoc analysis into reproducible dbt projects and iteratively fix issues, though human oversight and data quality checks remain important.
read more →

Kinesis Data Streams adds Dry Run API checks

🔍 Amazon Kinesis Data Streams introduces a DryRun option allowing customers to validate API requests and permissions without executing operations. Previously, teams used brittle workarounds like deliberately failing requests to check access, risking unintended writes if limits changed. The new optional DryRun parameter returns a DryRunOperationException when validation succeeds and is supported for five APIs across all Regions.
read more →

Cloudflare introduces OAuth scope customization

🔒 Cloudflare announced task-based OAuth consent to let client owners mark specific scopes as optional, enabling users to grant a narrower subset of requested access during authorization. The change builds on OAuth's existing ability to grant fewer scopes than requested and keeps existing client behavior by default. Developers must check granted scopes after the token exchange and can opt in when configuring a client.
read more →

Human‑Amplified AI for Security Research Advances

🔎 A new AI-driven system called HTTP Terminator found hundreds of live websites vulnerable to HTTP request smuggling and even proposed a novel class of flaw, “shared-parser confusion,” but it operated under continuous human guidance. PortSwigger researcher James Kettle designed the system around his own methodology, applying ideation, large-scale evaluation, anomaly detection, weaponization checks, and cascade analysis. Kettle open-sourced the tool and blueprint, stressing that human oversight, deterministic code and careful evaluation strategies amplified AI capabilities and produced more reliable, improvable research outcomes.
read more →

Prisma AIRS Integrates with OpenAI Codex

🔒 Palo Alto Networks announces native integration of Prisma AIRS Runtime API with OpenAI Codex, enabling centralized, API-level security controls for developer workflows. The integration inspects developer inputs and prevents sensitive data leakage without requiring client-side hooks, preserving developer productivity in Codex. SecOps benefit from consistent policy enforcement, audit-ready logging, and organization-wide visibility through the Codex Enterprise Management UI.
read more →

AWS WAF Adds Salt Security Managed Rule Group

🔒 AWS WAF now offers the Salt Security managed rule group, available via AWS Marketplace as Salt Managed Rules for AWS WAF - AI Agent & API Security. The group provides detection and mitigation for API-focused attacks and traffic from AI agents and Model Context Protocol (MCP) endpoints without requiring customers to write custom rules. It addresses threats such as credential brute force, excessive GraphQL queries, SSRF, prototype pollution, and JWT anomalies while labeling MCP traffic and sensitive request attributes. Customers can subscribe and add the rule group to a web ACL directly in the AWS WAF console; pricing and versioning are managed by Salt Security through AWS Marketplace.
read more →

MCP v2: Stateless Model Context Protocol Update

🛠️ The MCP 2026-07-28 specification makes the Model Context Protocol fully stateless, removing session handshakes and Mcp-Session-Id requirements. The update simplifies deployment by allowing MCP servers to run on request-scoped infrastructure like Cloudflare Workers, while preserving elicitation via Multi Round-Trip Requests and tightening authorization with CIMD and RFC 9207 issuer checks. SDKs and migration guidance accompany the release.
read more →

Deutsche Bank’s API-First Transformation with Apigee

🧩 Deutsche Bank adopted Google Cloud's Apigee to transform monolithic systems into a governed, scalable API ecosystem. The platform centralizes documentation, security policies, and governance while enabling discoverability and reuse across teams. Apigee enforces least-privilege access, rate limiting, and observability, and supports resilience, auto-scaling, and caching for high performance. This API-first foundation positions the bank for AI-ready, low-latency services and emerging standards.
read more →