Inspect a reproduction
Read the original source and evidence before starting an investigation.
Open the workspace
Open the URL printed after a successful capture. Sign in with an account that can access the reproduction's organization, or use the access granted by its sharing settings. A link alone does not grant all sandbox or editing permissions.
Confirm the title, failure description, and captured command or browser scenario. Compare them with the issue you intend to investigate.
Read the captured source
Browse the source files and open the files referenced by the failure. The bundle contains selected source, not every file on the original machine. Missing ignored files, credentials, or generated inputs can affect replay.
Command capture freezes source before executing the command. Later local edits do not silently become part of that original capture.
Inspect evidence
Read captured output and stack frames first. For browser captures, inspect available console and request diagnostics and any approved screenshot or video in Visual diagnostics.
If media is absent, check the capture's privacy choices. Recording locally does not imply the recording was approved for upload. Unreviewed media stays excluded.
Captured logs describe the original observation. They are different from output produced by a live sandbox terminal during investigation.
Decide the next step
Use the sandbox to explore source and run commands interactively. Use verification to request a fresh replay and compare its evidence with the captured baseline.
If the captured source or failure is wrong, make a corrected capture from your local application rather than assuming a sandbox session repairs the original evidence. See capture guides and troubleshooting.