Perfecto 26.8 now attaches an ai.txt file to execution reports when global knowledge is used in AI test flows, giving release owners concrete evidence of the context behind an AI-driven result. That evidence matters when teams need to explain a failed validation, approve a release exception, or demonstrate control over AI-assisted quality processes.
AI-assisted test flows gain an evidence trail
The new report attachment applies when global knowledge is used in AI User Action flows, including AI Scriptless and Selenium/Appium AI User Action implementations. If a test does not use global knowledge, no attachment is added.
For enterprise QA organizations, this changes the review posture from trusting an opaque AI interaction to examining the knowledge supplied to it. The attachment can be previewed or downloaded from the execution report, helping teams retain evidence with the same test record used for release decisions.
Additional AI Scriptless refinements reduce friction during test authoring and investigation:
- Test-flow steps can expand inline to show the full AI command.
- Variable suggestions appear when authors type
${in AI Validation and AI User Action fields. - Quantum 3.2.6 supports reusing values captured by earlier AI steps in later prompts, actions, and validations.
These changes support reuse and faster review, but teams should still define ownership for global knowledge, approval criteria for changes, and retention expectations for execution artifacts. The evidence make audit conversations more concrete.
DevTunnel renewal requires an operational response
Perfecto renewed DevTunnel certificates for affected cloud environments. A cloud may be affected when its DevTunnel certificate was issued before January 9, 2024.
Organizations that continue to use DevTunnel must reinstall it to obtain the updated certificate. An installation using the old certificate will not connect until reinstallation is complete. This is not a code change, but it can stop remote connectivity used by developer workstations and CI agents.
Release and platform owners should treat this as a controlled maintenance item:
- Identify clouds and agents that depend on DevTunnel.
- Confirm certificate issuance status with Perfecto Support if uncertain.
- Schedule reinstallations outside critical release windows.
- Validate connectivity from representative developer and pipeline environments.
- Record completion and any exceptions in the operational change log.
The renewed certificates have a longer validity period, reducing the likelihood of frequent repeat action.
API limits make integration behavior explicit
Perfecto now enforces rate limits for high-impact API operations, including file and artifact uploads. The default limit is 20 requests per 60 seconds. When a client exceeds that limit, it receives HTTP 429 and a Retry-After header that specifies when it can retry.
Automated test execution is not affected. The principal workflow impact falls on tools that upload assets, collect artifacts, or perform administrative operations at high volume.
Engineering teams should update API clients to respect Retry-After rather than retrying immediately. This protects shared platform capacity and prevents avoidable pipeline failures during high-concurrency activity. Where a documented workload needs a higher threshold, teams can request it through Perfecto Support.
Framework support updates dependency planning
Perfecto 26.8 adds support for Quantum 3.2.6, Quantum 3.2.5, and Playwright 1.60.0. Perfecto supports the latest Playwright release and the prior nine versions, giving teams a defined compatibility window for browser-test dependency management.
Quantum 3.2.6 makes Appium 2.19 the default for virtual-device tests when appiumVersion is not specified, alongside security and stability updates. Quantum 3.2.5 introduces configurable AI Visual Comparison behavior on failure: ignore is the default, while abort allows teams to stop execution when a visual comparison fails. It also corrects execution support for .bdd files.
Teams should make the visual-comparison setting an intentional policy decision. ignore may suit exploratory or signal-gathering suites; abort may better fit controlled release gates where an unresolved visual deviation must stop progression.
Mobile and browser coverage expands with build constraints
The release adds real-device coverage for iOS 27 Beta 4, iOS 26.6 GA, and iPadOS 27 Beta 3 version 2. Geolocation setting now works on iOS 27 real devices. Availability of iOS 27.x devices in a specific cloud requires coordination with Perfecto Support.
Virtual-device testing now includes iOS 27 Beta 4, with two important engineering implications:
- Tests targeting iOS 27 simulators must explicitly set
appiumVersionto3.3. - On iOS 26.4 and later virtual devices, applications compiled only for Intel x86_64 cannot run; teams need an ARM64 build.
Browser coverage now includes Chrome 151 Stable and 152 Beta, Firefox 153 Stable and 154 Beta, and Edge 151 Stable on Windows and Mac machines.
For release owners, beta operating systems and browsers should enter a separate compatibility lane before becoming a standard release gate. Validate application behavior, test-framework compatibility, device availability, and build architecture before changing the supported-platform commitment.
Recommended adoption sequence
A controlled adoption plan keeps infrastructure work from being confused with test failures:
- Complete DevTunnel impact assessment and reinstallation where required.
- Update shared API clients with 429 and
Retry-Afterhandling. - Review AI global-knowledge governance and report-retention practices.
- Pin and validate Quantum, Playwright, and Appium versions in representative pipelines.
- Verify ARM64 mobile builds and explicit Appium 3.3 capabilities for iOS 27 simulators.
- Add new browser and mobile versions through risk-based regression suites.
This sequencing gives teams a defensible record of readiness while maintaining delivery throughput.
How Merito helps
Merito helps enterprise teams turn release-note changes into an owned delivery plan. We assess DevTunnel and integration impact, define AI-test evidence and governance practices, update CI/CD test controls, and establish platform-compatibility rollout criteria. The result is a release adoption path that accounts for audit evidence, operational dependencies, and application risk.
Merito is an authorized Perforce reseller and services partner. Our Perforce Perfecto team can scope licensing, sizing, and rollout for 26.8, and our enterprise upgrade services help you plan and validate the upgrade with minimal disruption to release schedules.