FIELD NOTE · RELEASES · REMEDIATION
Preview, beta or GA: the release status of every AppSec agent we track
Short answer
By The Pipeline Desk · Best Agentic AppSec · Published 2026-10-06 · Updated 2026-10-11 · 4 min read
Why release status matters
A feature in preview can change, gain limits or disappear, and its support terms often differ from a generally available product. Our scoring credits generally available features more than previews, so the status of each feature matters for the scores as well as for buyers.
This list uses the status each vendor states on its own page. Where a page announces a launch without naming a stage, we write "Launched".
Generally available
- Snyk Evo Agentic Development Security. Announced 23 June 2026 to govern AI agents across the development lifecycle, with general availability set for 29 June 2026.
- Snyk Evo Continuous Offensive Security. Generally available from 4 August 2026, for AI-driven pentesting and red teaming of AI agents.
- Snyk Secrets. Generally available from 4 August 2026, with prevention gates across coding agents, IDEs, pull requests and CI/CD.
- GitHub agentic autofix for Code Quality findings. Generally available from 9 September 2026: up to 25 findings assigned to Copilot at once, fixed on a branch, validated and opened as a pull request.
Launched (stage not named on the page)
- Legit Security Agentic AppSec. Introduced 16 June 2026 and announced in a press release on 18 June 2026, with remediation agents that fix findings, re-scan and open pull requests; a later post extends the fixes to dependency vulnerabilities.
- Endor Labs AURI Agents. 11+ pre-built open-source agents, including AI SAST Remediation and SCA Remediation, free and open source for Endor Labs customers.
- OX Security AI-Powered Fixes. One-click fixes in CI/CD, IDEs and pull or merge requests.
- Legit Security VibeGuard 2.0. Launched 3 August 2026 for coding agents on the developer endpoint.
- OX Security VibeSec Dependency Vetting. Announced 9 September 2026.
- Endor Labs Threat Center and Package Firewall for VS Code extensions. Launched 9 September 2026.
- OX Cloud. Launched 16 September 2026 with AI Detection and Response and Agentic Attack Surface Management.
- Aikido Security Altar. Released 21 September 2026 as an open-weight security model.
- Cycode AI Guardrails. Launched 17 February 2026, with Cursor and Claude Code supported at launch.
- GitHub agentic autofix with Copilot Memory. From 25 September 2026, agentic autofix stores fix patterns in Copilot Memory for customers who have it enabled.
Public preview
- Snyk remediation agent. Public preview through the Snyk CLI and agentic development environments from 4 August 2026.
- GitHub agentic autofix for code scanning alerts. Public preview.
- GitHub merge blocking on open secret alerts. A ruleset rule in public preview from 9 September 2026.
- GitHub AI Scan for pull requests. Public preview; since 16 September 2026 it no longer requires CodeQL default setup, and it is not yet on GitHub Enterprise Server.
Public and private beta
- Semgrep Agentic Workflows. Nine workflows in public beta from 28 July 2026, covering 70+ CWEs.
- Semgrep Malware Firewall. Private beta, announced 24 September 2026 with malware detection and response automation for Semgrep Supply Chain.
What the list shows
Three patterns stand out. First, fix agents are the least mature group: one is generally available for a defined scope (GitHub, Code Quality), one is in public preview (Snyk), and the others, from Legit Security, Endor Labs, Cycode, OX Security and Aikido Security, are described without a stage. Second, governance of coding agents moved to general availability fastest, with Snyk's Agentic Development Security and Legit Security's VibeGuard 2.0 both shipping within six weeks of each other. Third, supply chain controls at install time arrived as a cluster in September, from Endor Labs, OX Security and Semgrep.
How to use it
When a vendor demo shows an agent feature, ask for its stage and the date it reached that stage, and check the plan it needs. Then compare it with this list. We will update the list when vendors change a status on their pages.
Questions to ask about a preview
- Is the preview open to every customer, or by request?
- Does it carry the same support terms and data handling commitments as the generally available product?
- Is there a stated date or condition for general availability?
- Will the pricing change when it leaves preview?
- Can it be switched off per repository if it misbehaves?
A preview is often the fastest way to see where a product is going. It is a weaker basis for a purchase that needs to hold for a contract term, which is why our scores give it partial credit.
Sources
Source: Snyk news: Evo Agentic Development Security launch · Read 27 Sep 2026
Source: Snyk news: Evo Continuous Offensive Security and Snyk Secrets · Read 27 Sep 2026
Source: GitHub Changelog: agentic autofix for Code Quality findings (9 Sep 2026) · Read 27 Sep 2026
Source: GitHub Changelog: agentic autofix uses Copilot Memory (25 Sep 2026) · Read 27 Sep 2026
Source: GitHub Changelog: AI Scan no longer requires CodeQL default setup (16 Sep 2026) · Read 27 Sep 2026
Source: GitHub Changelog: block pull requests with exposed secrets (9 Sep 2026) · Read 27 Sep 2026
Source: Semgrep blog: agentic workflows · Read 27 Sep 2026
Source: Semgrep blog: malware detection and response automation · Read 27 Sep 2026
Source: Legit Security blog: Agentic AppSec · Read 27 Sep 2026
Source: Legit Security: VibeGuard 2.0 press release · Read 27 Sep 2026
Source: OX Security blog: VibeSec Dependency Vetting · Read 27 Sep 2026
Source: Legit Security press release, 18 June 2026 · Read 11 Oct 2026
Source: Legit Security blog: Agentic Remediation now fixes dependencies · Read 11 Oct 2026
Source: Endor Labs: AURI agentic workflows · Read 11 Oct 2026
Source: OX Security: AI-Powered Fixes · Read 11 Oct 2026
Source: OX Security blog: OX Cloud for AI agents · Read 27 Sep 2026
Source: Endor Labs: Threat Center and Package Firewall for VS Code extensions · Read 27 Sep 2026
Source: Aikido Security blog: Altar open-weight model · Read 27 Sep 2026
Source: Cycode blog: AI Guardrails in the IDE · Read 27 Sep 2026
Read next
Lesson 9: Reading vendor agent claims
How to read vendor pages about AppSec agents: release status, what 'validated' means, which agents are named, which plan includes the feature, and what is not stated.
Agentic AppSec rankings
The top 10 platforms ranked on nine weighted criteria.
Reading the ranking
Why the top of our agentic AppSec ranking looks the way it does, which criteria decide it, and where the leader trails, computed from the published scores.
More field notes
2026-09-30 · 4 min read
How six vendors enforce coding agent policy: endpoint, IDE, hook or guidance
Six vendors, four enforcement models for coding agent policy: endpoint control, IDE guardrails, hooks inside the agent, and guidance before code is written. What each documents.
2026-10-04 · 3 min read
Pentest agents and agentic workflows: where offensive AI fits next to AppSec platforms
Snyk Evo Continuous Offensive Security, OX Agentic Pentester, Semgrep agentic workflows and Aikido Altar: what each vendor says its offensive AI does, and how it relates to triage.