data/sessions/explorer-extract-zip/memory.md

Memory: explorer-extract-zip

Overview

Extract a zip archive in place through the file manager's extraction wizard. Use this when an archive's contents are wanted beside the archive. The route right-clicks the archive and chooses Extract All, then accepts the destination the wizard has pre-filled — which is a folder named after the archive, in the same place. The recorded trace extracts Archive.zip beside itself, leaving the archive intact.

Prerequisites

• A desktop environment with a file manager that can expand archives, or a shell with an extraction tool.
• An archive file that exists and is readable, in a format the tool handles.
• Write access to the destination folder, and enough free space for the uncompressed contents — which can be far larger than the archive.
• Knowledge of whether the archive contains a single top-level folder or a set of loose files. This determines whether extracting "in place" produces one tidy folder or scatters files across the destination.
• No existing files at the destination that the archive's contents would replace.
• An archive from a source that is trusted enough to unpack. Extraction writes whatever paths the archive names.

Steps

Step 1 — Move-cursor at (250, 238)

The task is to unpack Archive.zip into a folder. Move the cursor onto its row in corpus-seed first, without clicking — the move-then-click pair is what makes the click land reliably here.

Step 2 — Right-click at (250, 238)

Move the cursor and Right-click Archive.zip to open its context menu. A double-click would only browse the archive's contents in place, which reads files out of it but never writes them to disk; extracting needs the explicit command on this menu.

Step 3 — Move-cursor at (316, 348)

Move the cursor onto the Extract All... row, letting the hover highlight confirm the target before committing.

Step 4 — Left-click at (316, 348)

Move the cursor and Left-click Extract All.... This opens the extraction wizard rather than extracting immediately, so the destination can be reviewed before anything is written.

Step 5 — Move-cursor at (1140, 722)

The wizard has pre-filled the destination as C:\Users\AgentUser\corpus-seed\Archive — a folder named after the archive, next to it — which is exactly the layout the task wants, so nothing needs to be typed or browsed for. Move the cursor onto the Extract button.

Step 6 — Left-click at (1140, 722)

Move the cursor and Left-click Extract, accepting the default destination. The archive's contents are written into the new Archive\ folder, and because Show extracted files when complete is ticked, Explorer opens that folder afterwards — which doubles as the confirmation that the extraction ran.

Notes & Edge Cases

1. An archive opens as a browsable folder on double-click on this system, which is why extracting needs the menu entry instead. Browsing an archive and extracting it are different things, and only one of them leaves files on disk.
2. The entry opens a dialog with the destination already filled in from the archive's name. That destination is a decision the dialog has made rather than one the task made, so read it before confirming rather than afterwards.
3. The dialog's checkbox only decides whether the extracted folder is opened at the end. It changes what is in front afterwards and nothing about the result, so it is not the confirmation it looks like.
4. The archive is not consumed. The folder ends with the archive and its contents side by side.
5. A destination that already exists is merged into rather than replaced, and same-named files inside prompt one at a time. A run expected to be unattended can stall on that prompt indefinitely.
6. Free space is measured against the uncompressed size, which the archive's own size says nothing useful about.
7. An archive from an unknown source can name paths outside the destination. Inspect the entries first when the source is not trusted.