mabl adds agent guardrails and terminal checks for AI-written code
mabl, the agentic testing platform, has introduced capabilities that give engineering and quality teams an independent check on software produced by coding agents. According to the company, teams can set the standards their agents are held to and see the impact of a change from the terminal before a pull request.
- mabl introduced agent guardrails, a knowledge graph and terminal-based impact checks for code written by coding agents.
- Test cases in Xray, Zephyr, TestRail and Azure DevOps serve as the standard agents are held to, with results flowing back.
- Desktop application testing is coming soon for enterprises.
- In 90 days, coding agents made 3.0 million MCP tool calls to mabl and created more than 4,200 tests, according to the company.
- mabl estimates it saved customers about 100,000 engineering hours.
What mabl released
The release, announced from Boston, targets what mabl describes as a verification problem created by coding agents. The company says coding agents are changing how software gets written, and that mabl provides an independent check on that work, grounded in how the application behaves and built to keep pace with every change.
The launch accompanies an open letter from mabl co-founder Dan Belcher titled "Testing takes center stage in the ADLC". Belcher said in the announcement: "Our customers want to move faster with coding agents without losing trust in what ships." He added that the release "puts mabl's independent verdict inside the developer's workflow and gives quality teams the controls to define what good looks like."
The new capabilities
According to mabl, the following are available today:
- Agent guardrails: teams set objectives that keep tests tied to their goals, and guardrails that define what agents may do on their behalf, including coding agents that call mabl.
- The mabl knowledge graph: when a change is in progress, mabl assesses its impact, finds the tests that verify it and highlights coverage gaps.
- Verification in the builder's workflow: mabl works alongside the coding agent, so developers can see the impact of a change from their terminal before opening a pull request.
- Suite management at scale: users can ask mabl to coordinate work across workspaces, such as repetitive maintenance.
- Test case management integrations: test cases in Xray, Zephyr, TestRail and Azure DevOps become the bar every agent is held to, and results flow back to those tools.
Desktop application testing is also part of the announcement, but it is listed as coming soon for enterprises modernizing their stacks. It extends independent verification to desktop applications.
Usage figures from mabl
The company shared usage numbers for the last 90 days. mabl's agents analyzed 1.58 million unique failures across 3.3 million agent sessions. Coding agents made 3.0 million MCP tool calls to mabl and created more than 4,200 tests. mabl estimates that this saved customers about 100,000 engineering hours. These figures come from the company itself and are not independently verified in the announcement.
What it means for teams
For engineering teams, the main change is where verification happens. Instead of waiting for a review stage, developers can check the impact of an agent's change in the terminal before they open a pull request. For quality teams, the guardrails and the link to existing test case tools such as Xray, Zephyr, TestRail and Azure DevOps are meant to let them define the standard that every agent is measured against.
mabl says the release will be introduced at mabl Experience 2026 in Boston. The company points interested users to the open letter, its documentation for launch details, and a free trial.
