The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →To audit an AI assistant, compare its file changes with a known baseline, review the session record, and check any external services it could access. To undo work, use the recovery method that matches the effect: an IDE review or checkpoint for supported workspace edits, Git for repository history, and the affected service’s own controls for commands, deployments, or API changes. No single undo button covers all three.
Before the assistant acts, establish what it can change
For repository work, check the working tree first. A clean baseline makes it easier to distinguish assistant changes from existing work; if there are already edits, identify them before proceeding. Decide which files, tools, terminal commands, networks, and external services the task requires, and scope permissions and approval rules to that work wherever the host allows it.
For untrusted projects, Microsoft recommends using VS Code Restricted Mode. Its security guidance also recommends agent sandboxing on supported platforms, reviewing edits before integrating them, protecting sensitive files, and keeping auto-approval scoped to a session. Commands run with your credentials can push code, change infrastructure, call APIs, or trigger deployments. Read Microsoft’s VS Code security guidance.
Review what changed and preserve a useful trail
- Inspect the changed-file list and diff. Look for additions, deletions, and edits to sensitive configuration, then compare meaningful changes with the task you gave the assistant. A diff shows file changes; by itself, it does not record every command, approval, network request, or external effect.
- Keep a stable repository reference. Use Git for durable history and collaboration. Record the session identifier or retain the relevant session log as well: a commit identifies the code state, while a session record may help explain how it came about. Git’s Pro Git book covers version control fundamentals.
- Capture activity beyond file edits when it matters. Keep relevant commands, tool invocations, approvals, timestamps, changed paths, resulting commits, and external services touched. In VS Code, agent hooks can record tool invocations, command execution, and file changes. GitHub’s Copilot SDK documentation describes lifecycle hooks for custom safety checks, audit logging, and approval workflows; logging should be scoped and protected, and hooks should be tested in the host where they will run. See VS Code security guidance and GitHub’s Copilot Agents documentation.
In VS Code, supported sessions let you review changes before commit, merge, or pull request. Checkpoints capture affected workspace files before a request, and the interface can show which files and how many lines changed. In the extension-host workflow, pending edits can be accepted or rejected individually, or resolved from the chat view. See Microsoft’s review and revert documentation.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minute#1 Best Overall
For GitHub Copilot cloud agents, session pages show progress, token usage, and session length. Cloud-agent commit messages link to session logs, which can help trace a change during review or an audit. See GitHub’s session-management documentation.
Choose a recovery method that matches the change
| Recovery method | Useful for | What it can restore or explain | Important limit |
|---|---|---|---|
| IDE diff or pending-edit review | Inspecting edits before accepting them | Which files and lines changed; supported edits can be accepted or rejected | Host and workflow vary, and changes may already be written to disk before review. VS Code documentation. |
| IDE checkpoint | Returning supported workspace files to an earlier request state | Affected files and, in VS Code, subsequent chat history | Temporary; does not undo completed commands or external effects. VS Code documentation. |
| Git history | Durable repository history, collaboration, and code recovery | Tracked file states and committed changes | Does not by itself explain the assistant’s intent or reverse effects in external services. Pro Git. |
| Agent session log | Reconstructing a session | Depending on the platform, prompts, responses, progress, file changes, or linked commits | Availability, content, access, and retention vary. GitHub session documentation. |
| Enterprise audit log or compliance API | Organization-level activity investigation | Administrative or agent events, filters, and exports | May omit chat content; access, plan eligibility, retention, and export limits vary. See GitHub, OpenAI, and Anthropic. |
| External service audit and recovery controls | Actions against APIs, cloud services, deployments, and other systems | Whether an outside action occurred and what recovery that service supports | Separate from workspace rollback; identify each affected service and use its own controls. |
Undo supported workspace edits in VS Code
- Open the earlier request in the supported chat session.
- Choose Restore Checkpoint.
- Review the restored files and conversation state before continuing.
Restoring a checkpoint also removes subsequent requests from the conversation history. In Microsoft’s words, “A checkpoint restores affected workspace files and chat history. It doesn’t reverse completed terminal commands, network requests, deployments, or changes that tools made to external services.” Microsoft also states: “Checkpoints are temporary and don’t replace Git version control.” Both statements are from its review and revert documentation.
Recover Git-tracked changes without discarding unrelated work
First inspect the working tree and history, then match the recovery operation to the change’s state. For unwanted unstaged or staged edits, restore only the files or hunks you have identified as unwanted. For a committed change, consider a revert commit when the change is in shared history. The right operation depends on whether the change is unstaged, staged, committed locally, or pushed; avoid treating a branch reset or deletion of unreviewed work as a universal rollback.
Recover actions outside the workspace separately
A file checkpoint cannot reverse a completed terminal command, network request, deployment, or external-service change. Stop a still-running agent if necessary, then inspect each affected service directly. Check its audit events, version history, transaction controls, backups, or recovery procedure; determine whether the operation completed or was repeated, and consider whether a compensating action could cause further harm. If code was pushed, identify the commit and coordinate repository recovery with the team.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #3
Stopping an agent is not the same as rolling back its work: GitHub says stopping a Copilot cloud-agent session ends the Actions run but preserves commits already pushed. Review those commits and use repository history if recovery is needed. See GitHub’s session-management documentation.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Know what the logs contain—and how long they last
A session transcript and an administrative audit log serve different purposes. A session record may help reconstruct prompts, responses, and file changes; an administrative log may focus on events and omit message content. Before relying on a provider’s records, check the eligible plan or role, event coverage, export method, content detail, and retention period. The published periods below describe different products and record types, not comparable measures of assistant reliability.
Rank #4
| Provider and record | Published retention | Scope and qualification |
|---|---|---|
| GitHub Copilot enterprise audit log | 180 days | GitHub’s Enterprise Cloud documentation, accessed October 4, 2026, says the enterprise audit log retains events for the last 180 days and recommends streaming events to a SIEM for longer history. GitHub audit-log documentation. |
| OpenAI Compliance Logs Platform | 30 days | OpenAI’s documentation, accessed October 4, 2026, says the platform retains data for 30 days and advises customers needing longer retention to download and retain logs continuously under their own policies. It also notes a stateful route deprecation and removal in 2026, so API users should check the current API reference. OpenAI Compliance Platform documentation. |
| Anthropic Enterprise audit-log export | 180 days | Anthropic’s “Access audit logs,” dated June 15, 2026, covers the organization’s previous 180 days. The audit logs do not include chat titles or content; chat inputs and outputs may be available separately to Primary Owners through data exports. Anthropic audit-log documentation. |
Log access and detail are not interchangeable across plans or providers. Audit data can expose sensitive operational information, so restrict access and choose an export and retention policy suited to the organization’s needs.
Quick Recap
Best Value
Reduce the chance and impact of unwanted actions
- Keep write permissions, tool access, and command approvals limited to what the task needs.
- Use sandboxing and network restrictions where supported; they reduce exposure but do not make completed external actions reversible.
- Require human review before integrating consequential edits or approving high-impact actions.
- Protect sensitive files and avoid broad auto-approval. VS Code warns that rule-based terminal auto-approval relies on best-effort command parsing and has limitations; model-assisted permissions can make mistakes and are not a security boundary. See the security guidance.
- For Copilot CLI, GitHub describes a permission prompt as “An interactive confirmation step in Copilot CLI that asks the user to approve an action—such as modifying a file, executing a command, or accessing files outside the current directory—before the agent proceeds.” Treat prompts as one approval layer, not a substitute for reviewing changes and limiting access. GitHub Docs: GitHub Copilot Agents.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
