Repository navigation
·
3 commits
to main
since this release
Highlights
Scroll actions now fail loudly instead of silently doing nothing, and kane-cli reports exactly what is scrollable on a page — including inside cross-origin iframes and dialogs. Remote runs get first-class variable and secret handling, with per-member pool variables and vault-backed secrets, and coverage reporting now resolves straight from your Test Manager project without extra flags.
[0.10.6] - 2026-11-17
Scrolling that tells you the truth
- A scroll that moves nothing now fails — authoring scrolls, coordinate (vision) scrolls, and scrolls that never ran are reported as failures, the same way a locator click that misses one is, instead of passing quietly.
- Until-scrolls fail when the condition is never met — no more runs that scroll forever and then move on as if they succeeded.
- Scroll detection reaches further — scrollable regions inside cross-origin (out-of-process) iframes, dialogs in front of the page, and
<body>scrolling under a locked<html>are all detected and tagged, including horizontal scrollers. - Text-mode page output shows what scrolls — the page document lists scrollable containers, so you can see what the agent has to work with.
- Scroll results are judged on scroll offsets — the screenshot comparison is gone, so a visually similar page no longer hides a scroll that never happened.
Remote runs handle variables and secrets properly
- Pool variables per member — a
--remotejob names each member's pool variables and carries their secrets from the vault, so remote runs behave like local ones. - Secrets keep their bare names — remote member secrets are pushed under the same plain names your laptop runs use, stored under a vault key kane-cli owns.
- Global vs. secret is decided by your authoring run — on the grid, the authoring record determines which bare names bind as globals and which as secrets, and bindings are carried through replays.
- Empty seeded globals resolve correctly — a global seeded with no value on the grid now resolves through the binding instead of dropping out.
- Variable scoping is clearer — variables are step-scoped, environment names follow one rule, and globals stay laptop-only.
Coverage without the extra flags
- Coverage comes from your Test Manager project — the final evidence merge and the pack shipped by a task ask the project for coverage directly; no
--coverage-projectflag needed in a Test Manager job. --remotejobs name the project by id — Test Manager job output identifies its project by id, and both the flag and the job skip a project with no use cases.- Unknown environments are rejected — the cloud seal's coverage step refuses an unknown
--env, refuses a non-coverage answer, and refuses another environment's profile. - Logs stay clean — job login details and failure details stay out of the printed log (the reason for a failure is kept in the private log), and coverage comments and test names no longer carry internal triage ids.
Mobile replay of API responses
- Stored API response values replay on mobile — a mobile
api_responseextraction keeps its Python extraction code, replay carries the extracted outputs through to the binding, and hand-off follows the tape's platform. - Missing extraction code is named — replay tells you which extraction is missing its code instead of failing opaquely.
Fixes and defaults
- Self-healing runs on every web replay — page-code healing goes through the analyze loop on each web replay, and heals now warn and bind rather than silently rewriting.
- Dropdown options stay targeted — when selector resolution retargets to a parent
<select>, the original option locator is preserved. - Local dev serves its own bundle — the dev controller serves the bundle directly; no storage account required.
- Pinned dependencies — testmuai bindings
0.1.59b2and code-gen-framework>=1.3.6.
Install
# npm
npm install -g @testmuai/kane-cli@0.10.6
# Homebrew
brew install lambdatest/kane/kane-cli
# Shell installer (latest release)
curl -fsSL https://raw.githubusercontent.com/LambdaTest/kane-cli/main/install.sh | bashCompare: 0.10.5...0.10.6