Aikido Release Notes

Deep Review is Aikido's agentic reviewer for pull requests. It reasons across your full codebase and the discussion already happening on the PR, then posts findings and suggested fixes directly on the pull request, the way a senior security engineer would if they had time to read every change.

What it does

• Reviews with full context. Sees what you've discussed or flagged on the PR, plus what your other checks already flagged, not just the diff.
• Reasons across repos. It traces logic across files, including middleware and global config, and pulls in the linked repositories a change depends on.
• Posts fixes, not just comments. Findings land on the PR with a suggested fix attached.
• Saves feedback to memory. Mark a finding as an accepted risk or a false positive, and Deep Review will remember that and learn from it.
• Closes its own findings. When you push a fix, Deep Review confirms it actually resolves the issue, then resolves the comment for you.

What it catches

Deep Review flags issues that come from how a change interacts with the rest of your app, the type that usually surfaces in a pentest or in production:

• Deeply nested security issues
• Broken access control
• Business logic flaws
• Cross-tenant data leaks

Security is just part of it. Deep Review also flags bugs, unintended side effects, and suggests concrete improvements on every pull request.

How it fits into Aikido

• Runs after your other PR checks (like SAST and dependency checks) so it can build on their findings instead of repeating them.
• Shifts deep security review left in your development workflow, catching before merge what a pentest would only find later. It’s the continuous complement to your one-off Code Security Audit.

Deep Review is enabled per repository so it reasons across your whole codebase, and you can set the review level that fits your team.

How much does it cost?

Deep Review runs on credits from your Aikido Wallet: 1 credit ($1) per PR, with unlimited re-reviews on the same PR. This makes it 5-10x more accessible than the same review done in-house or with competitors. Get an estimate for how much it'll cost across your repos, and set a monthly cap to keep spend predictable.

→ Read more about Deep Review

Features

• Device Protection for Linux - block malware at install time on Linux desktops and servers, now in beta. Ships as .deb and .rpm
• GCP VM Scanning - Google Cloud VMs are now live for everyone. Agentless set-up, no SSH keys or agents. AWS, Azure and GCP all covered.

Aikido Intel - Malware & Vulnerability Database

Our research team has discovered over #371 new vulnerabilities over the last two weeks!

We burned 11.7bn tokens for you, just to find the best cyber AI model. We gave 10 models 32 fresh CVEs to rediscover, three passes each, and open weights just beat the public frontier.

But Aikido still does it better: we put Code Security Audit head to head with Claude Security on Mythos. 68 of 89 vulns found vs 60, at $75 vs $157. The harness beats the model.

Improvements

• General - Request to ignore issue on PR/MR with admin approval
• General - Ignore license issues by path
• CVE Exploitability - Support for containers
• Reports - OSCAL FedRAMP DAST report
• Package Health - Support for Nuget and Jetbrains
• Zen Firewall - Agent releases for Dotnet, Java, PHP, Ruby, Golang and Node
• IDE - New release for VSCode
• API - Bulk update Github PR configuration
• API - Request a hardened container image
• API - List full repository code quality findings

Last updated 7 days ago

Features

• Deep Review - Agentic PR review that goes beyond the diff, diving into the surrounding code and linked repositories to catch business logic mistakes, cross-service security issues and more
• Android Pentest - Aikido Pentest now covers native and hybrid Android apps, including those with biometric login
• Custom Reports - Describe the report you want in plain language, Aikido Agent builds it from your findings, and you share it with your workspace in a click

Free Credits!

Need credits for Deep Review, Security Audit? Pro and Advanced plans now get 100–200 free Aikido Credits every month! Keep using them and they keep coming.

Aikido Intel - Malware & Vulnerability Database

Our research team has discovered over #510 new vulnerabilities over the last two weeks!

Shai-Hulud: Part Three, a compromised keyv maintainer account seeded a self-replicating worm that hit 444+ npm packages and 2 billion monthly installs. Aikido pulled it apart, payload and IOCs included.

Improvements

• Reports - Custom Reports
• Package Health - Added support for PyPI and Nuget
• Cloud - Kubernetes team annotations are now surfaced as container labels
• Cloud - Kubernetes agents now have FIPS variants
• Integrations - Aikido Terraform provider
• Integrations - Gitlab now supports resolving of issues
• Integrations - Bypass PR checks per repository in Training mode
• Reports - Repository filter for trends over time
• Zen Firewall - Agent releases for PHP, and Ruby
• IDE - New releases for VSCode, and Visual Studio
• API - Read and update AutoFix settings

Last updated 1 day ago

Aikido's Data Exposure Audit is a code-based DSPM. It helps developers and AppSec teams find sensitive data exposure risks by analyzing application code and related files already connected to Aikido. It uses LLM agents to read your code, mapping out how sensitive data is created, moved, stored, and exposed. It tells developers which code paths create risky data exposure without accessing your production data.

What it does

• Finds sensitive data exposure before production. Identify risky data handling in code before sensitive data lands somewhere it shouldn’t.
• Maps sensitive data flows. Follows sensitive fields from sources like request bodies, uploads, DB reads, config, and application logic into sinks like logs, third party APIs, LLM prompts, caches, queues, and storage.
• Helps identify access risk. Shows where routes, roles, services, and tenant checks appear to control access to sensitive data.
• Finds exposures traditional storage analysis can miss. Unsafe logging, broad API responses, PII in vendor payloads, secrets in frontend bundles, and sensitive data sent to AI systems.

What it looks for

• Credentials in URLs. Tokens and API keys in query strings, redirects, and clone URLs, where logs, proxies, and browser history pick them up.
• API over-exposure. Endpoints returning more fields than the caller is entitled to, and missing tenant ownership checks on data access.
• Unsafe logging and exports. Unredacted PII and secrets in application logs, traces, caches, and dumps.
• Over-broad third-party and AI sharing. Customer data and secrets sent to vendors, analytics, LLM prompts, and vector stores beyond what's required.
• Weak storage. Regulated data stored unencrypted, or with base64, MD5, and SHA-1 standing in for encryption.
• Incomplete deletion. Delete flows that leave downstream records, caches, and related stores intact.

Plus shadow copies in staging and snapshots, over-permissioned IAM roles and OAuth scopes, and unauthenticated data stores.

From finding to fix

• Every finding includes code evidence, a data flow explanation, the exposure risk, and remediation guidance.
• AutoFix opens a fix PR where the remediation is code-level.
• Re-test any finding to confirm the issue is resolved.
• Runs against the repos you've already connected. No new data source to onboard.

Each Data Exposure Audit uses credits from your Aikido Wallet.

→ Read more about code-based DSPM

Aikido's Data Exposure Audit is a code-based DSPM. It helps developers and AppSec teams find sensitive data exposure risks by analyzing application code and related files already connected to Aikido. It uses LLM agents to read your code, mapping out how sensitive data is created, moved, stored, and exposed. It tells developers which code paths create risky data exposure without accessing your production data.

What it does

• Finds sensitive data exposure before production. Identify risky data handling in code before sensitive data lands somewhere it shouldn’t.
• Maps sensitive data flows. Follows sensitive fields from sources like request bodies, uploads, DB reads, config, and application logic into sinks like logs, third party APIs, LLM prompts, caches, queues, and storage.
• Helps identify access risk. Shows where routes, roles, services, and tenant checks appear to control access to sensitive data.
• Finds exposures traditional storage analysis can miss. Unsafe logging, broad API responses, PII in vendor payloads, secrets in frontend bundles, and sensitive data sent to AI systems.

What it looks for

• Credentials in URLs. Tokens and API keys in query strings, redirects, and clone URLs, where logs, proxies, and browser history pick them up.
• API over-exposure. Endpoints returning more fields than the caller is entitled to, and missing tenant ownership checks on data access.
• Unsafe logging and exports. Unredacted PII and secrets in application logs, traces, caches, and dumps.
• Over-broad third-party and AI sharing. Customer data and secrets sent to vendors, analytics, LLM prompts, and vector stores beyond what's required.
• Weak storage. Regulated data stored unencrypted, or with base64, MD5, and SHA-1 standing in for encryption.
• Incomplete deletion. Delete flows that leave downstream records, caches, and related stores intact.

Plus shadow copies in staging and snapshots, over-permissioned IAM roles and OAuth scopes, and unauthenticated data stores.

From finding to fix

• Every finding includes code evidence, a data flow explanation, the exposure risk, and remediation guidance.
• AutoFix opens a fix PR where the remediation is code-level.
• Re-test any finding to confirm the issue is resolved.
• Runs against the repos you've already connected. No new data source to onboard.

Each Data Exposure Audit uses credits from your Aikido Wallet.

→ Read more about code-based DSPM

Features

• Data Exposure Audit (DSPM) - Track how sensitive data flows and where it gets exposed, read entirely from your code. No database connections, no data access to grant, just point it at your repo and fix the exposures at the source.
• Aikido Libraries - Fix open-source CVEs on the exact version you already run. No major-version bumps, no breaking changes, no refactor to schedule, only a PR you merge.
• Code Quality Repository Scans - Find code quality issues that already exist in your repositories. Use them to find and prioritize technical debt, then keep track of your results as your code evolves.

Aikido Intel - Malware & Vulnerability Database

Our research team has discovered over #298 new vulnerabilities over the last two weeks!

We dug into wp2shell, the unauthenticated RCE in WordPress core, read the story here. Our Attack Surface scan detects it against your exposed assets automatically, so set it up to catch any unpatched WordPress in your estate

Improvements

• General - Aikido Inbox: Admins can review ignore and snooze requests from their team
• Cloud - Noise Reduction for Container Images
• General - Bitbucket server support for PR Checks
• Cloud - K8s registry option for public image scanning
• Pentest - Reduced pricing for tiny apps
• Code Quality - PR/MR Gating by Severity
• Code Quality - Auto Resolve findings on GitHub
• Code Quality - New rule: "Keep the diff minimal"
• Device Protection - Support for Jetbrains
• Zen Firewall - Agent releases for PHP, Ruby, Golang, and Node
• IDE - New releases for VSCode, Jetbrains, and Visual Studio
• Intel - Health score for Maven ecosystem
• Settings - Bulk assignment of clouds to teams

Last updated 7 days ago

Features

• Code Quality Repository Scan (early access) - Scan your whole codebase for quality issues, not just new PRs, monitor trends and spot the hotspots worth refactoring first
• Approval Inbox for Ignores - Route every ignore request through one inbox for admin sign-off, so nothing leaves your feed without a decision.
• Noise Reduction for Secrets & CVE’s - New AI layer on top of your rules checks each secret and CVE against your code to flag what's truly exploitable and quiet what isn't

Aikido Intel - Malware & Vulnerability Database

Our research team has discovered over #245 new vulnerabilities over the last two weeks! Watch the live feed → intel.aikido.dev

Improvements

• General - New Dashboard for all users
• Source Code - Azure DevOps Self-Managed Server support
• Integrations - Added ticket support for Github Server and Azure DevOps
• Cloud
  • Added Render cloud misconfiguration scanning
  • K8s namespaces added to container labels
  • Terraform module for GCP project/organization onboarding
• Device Protection
  • Disable protection for up to 1 hour
  • Added PHP composer support
• Registry Proxy - Support for Google Cloud Artifact
• Zen Firewall - Agent releases for Java, Python, Ruby, Golang, and Node
  • Node - Support for ElysiaJS
• IDE - New releases for VSCode, and Jetbrains
  • Added Jetbrains CLion
• API - List snooze and ignored issues from PR
• API - Get Device Protection SBOM
• API - Get SLA settings
• API - List AutoFix history
• API - GDPR report
• API - Get, Add, Update, Remove Container label
• API - List repository PR configuration

Last updated 22 days ago

Aikido’s secrets detection and dependency engine now have LLM agents that use reasoning to identify truly exploitable issues.

• Secrets AutoTriage helps determine whether a detected value is likely a real credential and whether it creates actionable risk.
• CVE Exploitability Analysis determines whether a CVE is likely exploitable in your codebase and how urgently it should be addressed. From there, it can act on its own: upgrading, downgrading, snoozing, or ignoring findings automatically based on rules you control.

Each run is logged, so you can see what was analyzed and what actions the agent took.

Why it matters

• Verification on top of detection: LLM agents assess open-source dependency CVEs and leaked secrets against your codebase to see what is actually exploitable.
• Smarter prioritization: findings can be escalated or deprioritized based on real-world risk (e.g., live/high-scope secrets, practically exploitable reachable CVEs).
• Short rationale included: each outcome includes an explanation so teams can review and override when needed.

How these fits into Aikido

• Secrets AutoTriage: runs after entropy detection, determining whether a discovered secret is a real, exposed value and the riskiness. Most findings are handled by defined rules, but in ambiguous cases the surrounding code is used to determine what the credential is for, its liveness, and scope.
• CVE Exploitability Analysis: adds LLM reasoning based on how your code uses the package, then it can upgrade, deprioritize, or snooze/ignore the CVE automatically based on the codebase-aware findings.

→ Read more about Secrets AutoTriage

→ Read more about CVE Exploitability Analysis

Features

• Agentic Dependency AutoFix - Let Aikido resolve complex dependency interactions for you. Perfect for projects with many interlocking dependencies.
• Registry Proxy - Your private registry can now route through Aikido, blocking malware and risky-new packages and serving safe, patched versions before they hit your code.
• Device Protection - Now covers Ruby and Rust, blocking malware on install and adding gems and crates to your device SBOM.

Aikido Intel - Malware & Vulnerability Research

Our research team has discovered more than #232 new vulnerabilities over the last two weeks! We also scanned 1.2 million newly published packages for malware and found 3,066 new ones.

Separately, our new State of AI in Security & Development 2026 report surveyed 450 CISOs and developers on how AI is reshaping security, including why 1 in 5 hit a serious incident from AI-generated code.

Improvements

• AI Pentest
  • Letter of Attestation
  • Continuous Pentest adds Azure support
  • Continuous Pentest choose branches for 'On Schedule' runs
• Labels - Import labels directly from Github
• MCP - Release with new auth flow, workspace filters, branch filter and team filters
• Code Audit - Add additional instructions before an audit
• Zen Firewall - Agent releases for Dotnet, Python, PHP, Golang and Node
• IDE - New releases for Jetbrains
• API - Manage labels on repositories
• API - List container runners

Last updated 1 month ago

Features

• Threat models, visualized - Every AI pentest & code audit now ships a diagram of your app's architecture, roles, and trust boundaries, so you get the full picture at a glance.
• Device Protection - Now available for Windows, with MDM support to roll it out across your fleet.
• Labels - Attach any metadata you want to repositories and container images (business unit, PII, region, ... ) and filter you issue feed based on those labels.

Aikido Intel - Malware & Vulnerability Database

Our research team has discovered over #105 new vulnerabilities over the last two weeks!

The Miasma worm just moved its trigger out of package.json and into binding.gyp, a file almost no one reviews.

Learn how it works in our latest blog.

Improvements

• Code Audit - Retest issues after fixing code
• Endpoint - Export device SBOM
• Zen Firewall - Agent releases for Dotnet, PHP, Ruby, Golang, and Node
• Package Proxy - Malware scanning for your Private Registries
• Report - New CIS K8 Benchmark report
• MCP - Release 1.0.8 added new search repositories tool
• Task Trackers - Specify which repository tasks should be created for (applies to all task managers)
• API - List Zen users
• API - Add Exception, and List Devices, Activity Logs, Exceptions, Permissions Groups for Device Protection
• API - Export Single Repostiory report
• API - List notes
• API - Support for multiple task trackers

Last updated 1 month ago

Features

AI Code Audit - Pentest-grade reasoning on your source code. Catch logic flaws and other vulnerabilities that other scanners miss, just connect a repo.

Aikido MCP - Search and fetch your Aikido issues straight from your AI coding assistant. Ask for critical findings, leaked secrets, or open vulnerabilities in any repo and act on them with your favorite AI.

ServiceNow Integration - Connect ServiceNow as a task tracker and Aikido turns security findings into tickets automatically, routing them into the workflow your teams already use.

Aikido Intel - Malware & Vulnerability Database

Our research team has discovered over #365 new vulnerabilities over the last month!

Among them:

Google API keys keep working after you delete them. Our testing found a deleted key can keep authenticating for up to 23 minutes, long enough for an attacker with a leaked key to reach your data and APIs, including Gemini.

Improvements

General - Ask Aikido about CVE’s through chat

Code - Release gating: SLA Mode

Cloud

• Added support for Oracle Cloud with +50 rules
• Onboard Entire Alibaba organization
• VM Scanning - Added GCP support

Pentest

• Pause Pentest
• Monorepo support

Device Protection

• Get Inbox notifications in Slack or Microsoft Teams
• Added Golang ecosystem support
• macOS silent install for MDM

Zen Firewall

• Agent releases for Dotnet, Java, PHP, Ruby, Golang, and Node
• New view to check why an IP is getting blocked

IDE

• New releases for VSCode, Jetbrains, and Visual Studio
• Added support for IaC scanning in Jetbrains

API

• List Licenses
• List Malware
• Get IDE/MCP usage statistics

Last updated 2 months ago

Code Audit runs pentest-grade security reasoning directly on your source code, searching for logic flaws and other vulnerabilities that pattern-based scanners can't find.

Connect your codebases, confirm the price in credits, and start the audit. No staging URL, crawl, or pentest scope setup required.

Why Teams Use Code Audit

Reasoning, not pattern matching

Find hard-to-find bugs like cross-tenant data leakage that classical pattern matching misses.

10× cheaper than a pentest

Get pentest-depth reasoning across your codebase in minutes, not hours, and run it on demand.

Zero setup, just connect a repo

No staging environment, traffic to replay, or agents to deploy. Point it at your source and start the audit.

Mythos-ready defense

Defend against the kind of attacks frontier models now make trivial, with reasoning that matches what attackers can do.

How Code Audit Fits into Aikido

Code Audit is a sibling to Aikido Pentest. It uses the same agentic engine and issue experience, but stays fully codebase-focused.

Findings include code-based evidence, root cause analysis, remediation guidance, and AI-powered AutoFix for creating remediation PRs with a few clicks.

Pricing is shown upfront in credits before you start. Cost depends on the size and complexity of the selected codebases.

Use it when you want deep reasoning on source code without setting up a live environment. It works especially well for hard-to-test-by-URL codebases like:

• mobile apps,
• smart contracts,
• desktop apps,
• embedded software, and
• libraries.

→ Try out Code Audit

Features

Device Protection: Stop supply chain attacks at install time with Device Protection, a new agent that blocks malicious packages (npm, PyPI, Maven, NuGet, …) and risky IDE and browser extensions on your team's devices.

Aikido AI Chat is now embedded in AutoFix & Code Analysis Modal. Use the chat to interact in natural language and understand a security issue. Want to make a small change to the AutoFix PR? Just describe the change, without leaving the product.

Aikido Intel - Malware & Vulnerability Database

Our research team has discovered over #70 new vulnerabilities over the last two weeks! Intel flagged the latest PyTorch Lightning PyPI compromise, just the kind of threat Aikido Device Protection stops.

Improvements

Cloud - Support for Supabase common configuration issues

AutoFix - Add custom context to Autofix via .aikido files

Integrations - Automatically close Github issues on resolved issues

Aikido Infinite - Support for Custom Frequencies (ex: every 6 weeks)

Zen Firewall - Agent releases for Dotnet, Python, PHP, Ruby, Golang, and Node

Ruby added support for Outbound Request Blocking

IDE - New releases for VSCode, Jetbrains, and Eclipse

API - List Cloud rules

API - Get project / repository mapping

API - Get CIS and AWS CIS report

Device Protection is an agent that covers the full install surface of your developers and others their devices: malware blocking, extension governance, group-based policies, and a request-and-approval workflow. Designed for security while keeping developers productive.

What's included

• Real-time malware blocking powered by Aikido Intel, across npm, PyPI, Maven, NuGet...

• Extension governance for VS Code, OpenVSX, Chrome, and skills.sh

• Request and approval workflow: blocked installs trigger a one-click exception request reviewed inside Aikido

• Group-based policies: set different rules per team

• Continuous device inventory: always-current view of every package and extension per device

Developer Experience

The agent runs silently in the background. For the vast majority of installs, nothing happens and packages install as normal. Developers only see a notification when something is blocked or held for approval:

• Malware block: A desktop notification shows the package name and flags it as malicious.

• Policy block: A notification explains the package is restricted by your organization's policy. Developers can open the dashboard for details, but only an admin can change the policy.

• Approval required: If your policy requires approval for new packages, the install is paused and a request goes to the admin Inbox.

→ More information

Features

• Escalate a Pentest Vulnerability - Aikido Pentest Agents are cautious by default to avoid unintended impact. Removing agent restrictions enables deeper analysis, confirming either that additional protection layers hold up or that a vulnerability is genuinely exploitable before it's discovered by real attackers.
• Additional Context for AutoTriage - Provide Aikido's AutoTriage with specific knowledge about your environment to improve the accuracy of its findings and cut down on false positives.

Aikido Intel - Malware & Vulnerability Database

Our research team has discovered over #47 new vulnerabilities over the last two weeks! Including one that turned out far less scary than the headlines suggest: read our breakdown of the Axios CVE-2026-40175.

Improvements

• Reports - Trends Over Time now has an “out of SLA” graph
• Integrations - Directly reply or ignore via Gitlab MR comments
• Cloud - Reachability for containers is now available for Google Cloud Run
• AutoFix - Add custom context for AutoFix via .aikido file
• Pentest - Additional login methods: files, keys and scripts
• Integrations - Broker support for websockets
• Zen - Exclude specific users from rate limiting
• Safe-chain - Release 1.4.9 adds shrinkwrap to npm
• Zen Firewall - Agent releases for Dotnet, Golang, and Node
• IDE - New releases for VSCode, Jetbrains, and Visual Studio
• API - Add Azure container registry
• API - Add GCP registry
• API - Update headers for Domain scan

Last updated 3 months ago