web-test-reporter

This skill appears to implement the described testing and report-generation workflow, but review these points before installing or running it: 1) The SKILL.md expects the agent-browser CLI and the Python package python-docx, but the skill metadata did not declare required binaries/dependencies — ensure you have a trusted agent-browser binary installed and run 'pip install python-docx' in a controlled environment. 2) The instructions will ask you to provide project code paths and login credentials; never supply production credentials. Prefer a dedicated test account or sandbox environment and avoid giving broad access to source trees you don't trust. 3) Edit the build_report.py defaults (BASE and OUT) before running — it currently contains a hard-coded Windows user path (C:\Users\liuchao25...) that may be irrelevant or stale and could cause files to be written to unintended locations. 4) Confirm the agent-browser tool you use is trustworthy (it will control a browser and can interact with your application). 5) If you must provide source code for white-box testing, limit scope to the intended module and avoid uploading secrets; review code you give the agent for sensitive config or keys. 6) Run testing in an isolated environment (VM/container) where possible, and verify that the skill cleans up test data as promised. If you want higher assurance, ask the publisher for: a) a declared dependency list (agent-browser, python packages), b) confirmation the agent will not transmit credentials off-host, and c) a cleaned/parameterized build_report.py without hard-coded user paths.