Use VS Code’s Chat view to start a plan with /plan (or choose Plan in the agent-mode picker), turn the reviewed plan into an implementation, and verify the result. Use the Agents window as the task board when you need to assign or monitor several agent sessions across projects; use Chat beside the editor for concentrated work in one project.
What the VS Code AI agent task board does
VS Code has two related surfaces for agent work. The Chat view is attached to the project you are actively editing. It keeps the conversation, code context and immediate changes close to the editor. The Agents window is a broader monitoring board: it lists sessions, lets you select one, and exposes its conversation, files, changes and validation results without requiring you to switch among separate project windows.
As an Amazon Associate I earn from qualifying purchases.
The practical workflow is Plan → Implement → Verify → Review. Planning gives the agent a chance to inspect the repository and resolve requirements before it edits anything. Implementation applies the approved approach. Verification runs the relevant checks or exercises the application. Review is your inspection of the reported evidence and the actual diff.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Planning is not a guarantee that an agent will make the right change. It is a written checkpoint that makes scope, assumptions, affected files and verification visible before implementation. Microsoft describes the purpose this way: “Planning helps you clarify requirements, compare approaches, and identify risks before an agent changes your code.”
#1 Best Overall
- [Ultra-Lightweight Ear Buds Designed for Small Ears] Each earbud weighs only 3.7g and features a compact, ergonomic in-ear design made especially for small ears. Secure, low-profile, and comfortable for workouts, all-day wear or overnight sleeping.The ultra-slim profile sits flush in your ear with no hard edges poking or rubbing, so you can even sleep on your side without discomfort.
- [Immersive Stereo Sound with TOZO OrigX Technology] TOZO OrigX tuning delivers clear vocals, balanced mids, and natural stereo sound for music, podcasts, and videos.
- [Long Battery Life for Daily Use] Get up to 7 hours of playtime on a single charge, with up to 32 hours total using the charging case—ideal for workdays, commuting, and extended listening sessions.
- [Bluetooth 5.3 & Stable Connection] Bluetooth 5.3 provides fast pairing, stable wireless performance, and reduced dropouts as you move around home or office.
- [Deep Bass with Clear Vocals] High-performance drivers produce punchy bass while keeping vocals clean and detailed for everyday listening.
Start a coding plan in Chat
1. Open the right workspace
Open the VS Code project that contains the code, configuration and tests the agent must understand. A plan is only as useful as the workspace context available to the session, so begin in the repository where the change will be made rather than in an unrelated folder.
2. Choose the session target
Open Chat and select the appropriate agent session target when the picker offers a choice. The documented workflow supports Copilot sessions and can also use a Claude session where configured. The Agents window documentation also lists Codex with an existing ChatGPT sign-in and other BYOK configurations; availability depends on the account, extension and organization settings in your VS Code installation.
3. Invoke Plan mode
Enter /plan in Chat, followed by the outcome you want. You can instead choose Plan from the agent-mode picker. In Copilot sessions, planning may also be invoked automatically when you ask for a plan, but using the explicit command makes your intent unambiguous.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →A useful request has three parts:
- Outcome: what a user or system should be able to do when the work is complete.
- Constraints: compatibility requirements, files or APIs that must remain unchanged, performance or security boundaries, and any implementation preferences.
- Verification: the tests, build, lint command, manual scenario or other evidence that will demonstrate success.
For example: “Plan a change that adds CSV export to the existing reports screen. Keep the current JSON endpoint and browser support unchanged. Identify the files and dependencies involved, and specify the automated tests plus a manual check for a large report.” This asks for a result, limits the solution space and defines what “done” means without prematurely prescribing every edit.
Answer the agent’s questions before accepting its plan
The Plan agent researches the project and can ask clarifying questions. Treat those questions as a requirements review, not as an interruption to click through. If an assumption is wrong, correct it before the plan becomes an implementation instruction.
Questions worth resolving
- Which user flow or API is authoritative when several similar ones exist?
- Must the change preserve a public interface, database migration path or older runtime?
- Should the agent modify generated files, or only the source that produces them?
- Which test command is reliable in this repository, and are integration services required?
- What should happen for empty, malformed, very large or unauthorized input?
Ask the agent to explain trade-offs when it presents more than one approach. A short explanation of why it selected a module, dependency or test strategy can expose a hidden coupling before code changes make the decision expensive to reverse.
How to improve an incomplete plan
If the plan omits an affected area, state the missing requirement directly and ask for a revised plan. If the proposed verification is too weak, add the evidence you need: a named test suite, a type check, a build, a command-line exercise or a reproducible manual scenario. Do not approve a plan that says only “run tests” when the project has several test layers and the agent has not identified which one proves the requested behavior.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Review the plan before implementation
When planning finishes, VS Code presents a plan-review card. Read it as a change proposal. Check the scope, affected files, decisions, dependencies and verification steps against the repository and your request.
Rank #2
- 【IP68 Waterproof Multifunctinal Smart Watch】Torjalph fitness activity tracker smart watch is IP68 waterproof technology ,you can wear it when swimming,handwashing,it's rain-proof which means that you perform multiple exercise modes outside without worrying about sweat or rain corroding your fitness smart watch,and this watch equipped with a 230 mAh large capacity battery magnetic charging ,and this sports activity tracker watch can be used 5-7 days If fully charged.
- 【Larger 1.83'' Full Screen Watch & Multi-dial Selection】1.83 inch full HD color touch screen smart watch for men / women provides excellent 240*284 resolution image quality ,This Bluetooth smart watch offers a better touch experience and remarkable immersion and ,more than 200 watch faces for you to choose from App"Da Fit ", you can change the dial style according to your mood to wear keeping the world on your wrist
- 【Accurate Activity Tracker Smart Watch 】sports watch activity tracker uses sweat and dust resistant breathable silicone strap, our Fitness Tracker Smart Watch for Android iOS Phones supports 12 sports modes,walking,running,bicycle,swimming,badminton,basketball,football,rope skipping. This bluetooth calls smart watch for men women is lightweight and stylish, bringing more convenient and relaxing sports experience while also providing more protection.
- 【Your 24 Hour Personal Daily Health Monitor】Fitness watch for women men full touch screen IP67 waterproof bluetooth smartwatch for Android iOS Phones with heart rate blood pressure monitor multifunction Andriod smart watch with steps counter sleep monitor sports fitness watch, monitoring blood oxygen saturation allows you to understand your physical condition in real time to track fitness progress ,to know your health status in time.
- 【More Functions Bluetooth Call Smart Watch】This fitness activity smart watch full screen for iPhone and Android phones includes a bunch of new features including:The bluetooth smart watch can hang up the phone, receive and display WhatsAPP /Text /Facebook /Twitter /Messenger reminder.Support sedentary reminder, alarm clock, stopwatch, timer,weather information,music control, remote camera control compatible with Android 5.0 and above or iOS 9.0 and above smartphones.
Scope and affected files
Confirm that every listed file belongs to the requested change and that important files are not missing. A plan that edits a generated artifact without naming its source, or changes a shared utility without acknowledging its callers, needs clarification before implementation.
Decisions and dependencies
Look for choices that could change behavior outside the feature: a new package, a schema adjustment, a route change, a configuration flag or a migration. Make the agent state what it will not change when that boundary matters. If a dependency is optional, ask whether the existing stack can provide the same capability and what trade-off each option creates.
Verification steps
Verification should match the risk. A user-interface change may need a targeted test and a manual browser check. A parser or data transformation needs edge-case tests. A configuration change may require a build or startup check in addition to unit tests. The plan should name the commands or actions and describe the expected result, not merely promise that validation will occur.
Give feedback or select an action
The review card accepts overall feedback and, where available, inline feedback on individual plan items. Review the updated plan after the agent responds. Depending on the session, actions include:
| Action | When to use it | What it means |
|---|---|---|
| Implement Plan | The scope and checks are acceptable and you want the agent to edit the workspace. | Starts implementation using the reviewed plan. |
| Approve Plan Only | You want to record or share the approach before any code changes. | Approves the plan without starting implementation. |
| Implement with Autopilot | You accept the session’s automation behavior and its configured tool permissions. | Lets the agent proceed with the available automated actions; monitor it closely. |
| Reject | The request is wrong, incomplete or no longer needed. | Stops the plan from being implemented. |
Plan approval and approval of individual tool actions are separate decisions. Approving a plan does not mean every later terminal command, file edit or other tool action is automatically approved.
Implement deliberately, then verify the evidence
Start implementation only after the boundary is clear
Choose an implementation action from the plan-review card once the expected files and checks are clear. Keep the original outcome and constraints visible in the conversation so the agent can refer back to them while editing.
Ask for relevant checks, not a ceremonial test run
After the edits, ask the agent to run the tests that prove the requested behavior and to exercise the application where a test does not cover the full path. The useful output is evidence: command names, exit status, failures, and a description of what the check actually covered.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsIf a check fails, distinguish a product defect from an environment problem. Ask the agent to identify whether the failure is caused by the patch, a missing service, an unavailable secret, an unsupported runtime or an unrelated pre-existing failure. Record the distinction in the session rather than treating a green command with no meaningful coverage as proof.
Rank #3
- [Low Noise High Fidelity] New bee guitar cord with 22AWG oxygen-free-copper center conductors provides enhanced signal transfer, the PVC insulation material and the high-density stranded copper double shield provide noise rejection for quiet operation
- [Ideal for Choice] New bee 1/4 straight and right angle metal connectors, ensures a secure connection in any instrument, pedal, or amp jack. L shape great fro on stage guitar stands
- [Wide Compatible] Suitable for mixer, power amplifier, microphone, effector, decoder, equalizer, electoronic organ, electric guitar, electric bass and other electronic equipments. Widely used in the installation of smart home, karaoke speaker, professional stage audio, etc
- [Easy Disassemble] Making incidental repairs quick and easy! Very nice, flexible cable
- [Dont Hesite to Order] 24 hours customer services and professional technology team are standing by
Review the diff yourself
Read the changed files and compare them with the approved scope. Confirm error handling, authorization, data validation, compatibility and test quality. A passing test suite does not prove that an unrequested file was not changed, that a migration is safe, or that a new dependency is acceptable.
Open the Agents window as your monitoring board
Use the Agents window when you need a portfolio-style view of work assigned to sessions across projects. Microsoft’s guidance is to use it to track progress and review results in one place instead of switching between project windows.
Ways to open it
- Select the title-bar Open in Agents button.
- Run the Chat: Open Agents window command from the Command Palette.
- Run
code --agents. - Use the browser entry point https://insiders.vscode.dev/agents. The documented browser flow requires GitHub sign-in.
The browser and desktop experiences can differ by release and configuration. VS Code documentation marks Automations as a preview feature, so preview behavior and available controls can change.
Understand the board layout
The window contains a sessions list, a customizations panel, a chat area, a changes panel and a files panel. Select a session in the list to make it active. Use the chat area to read the agent’s reasoning and requests, the files view to inspect the workspace context, and the changes view to examine what the session actually modified.
Monitor a session from start to finish
1. Select the session, not just the project
Several sessions may work in the same or different repositories. Select the specific session whose status or output you need; the active selection determines which conversation, files and changes you are viewing.
2. Read Chat for intent and blockers
Chat shows the request, clarifying questions, plan, tool activity and reported verification. Look for unanswered questions, assumptions that were never confirmed and commands that were skipped after a failure. If the agent is waiting for a decision, answer it in that session so the decision remains attached to the work.
3. Use Files to check workspace context
The Files view shows the workspace associated with the session. Use it to confirm that the agent is operating in the intended project and to locate the files named in the plan. This is especially important when several repositories have similar folder names.
Free tools Windows power users keep installed
One-click scans. No signup required.
4. Use Changes to inspect the actual scope
The Changes view can show branch changes, uncommitted changes, all changes or last-turn changes. Choose the view that answers your question: last-turn changes help isolate the current request, while all changes show the broader state of the branch. Inspect the diff rather than relying on the agent’s summary.
Rank #4
- Digital Family Calendar: BSIMB digital calendar wall touch screen puts an end to your chaotic life. Designed for busy families, the digital calendar keeps you organised and in step with your family. Assign different colours and tasks to each family member to make family scheduling easier and more intuitive.
- Easy Setup and Auto-Sync: Plug in, connect to WLAN and auto-syn your electric calendar in minutes. Digital wall calendar compatible with Google calendar, iCloud calendar, Outlook calendar, Cozi calendar and Yahoo calendar. Add events directly on the device or through the ecalendar app. Electronic calendar stays current whether you're at work or traveling—no manual updates needed.
- All-In-One Home Assistant: Smart calendar builds habits and encourage independence with personalised chores. Display daily meal plan to reduce the daily ‘What's for dinner?’ stress. To-do lists keep track of everything you want. Turning chores into valuable achievements with stars and rewards, making life easier and more enjoyable.
- Manage Anytime Anywhere: BSIMB electronic calendar for family stay connected with the powerful ecalendar app. Access and manage your digital wall calendar anytime, anywhere with free features like stars & rewards, photo screensaver, chore charts, meals plans, to-do lists etc. Electronic planner paid features like Magic Import tailor advanced tools for users with special needs.
- Stay in Touch with Loved Ones: Share photos or videos from your phone to electronic wall calendar with app, instantly update your life moment, and keep in touch with your family, especially those who are far away; you can also surprise your smart calendar digital wall by sending him a wish card; it's also great for keeping tech-illiterate elders in touch with the youngsters.
5. Capture the validation result
Check which tests, tasks or terminal commands ran, whether they passed, and whether the output covers the requested scenario. If validation is absent, ask the agent to run it. If it is blocked, leave the session with a precise next action instead of marking the task complete.
Chat view or Agents window?
| Surface | Best use | What to monitor | Main trade-off |
|---|---|---|---|
| Chat view | Frequent editing, debugging or testing in one project. | The conversation, code context and immediate changes beside the editor. | It is less suited to a portfolio-style view across multiple projects. |
| Agents window | Assigning and monitoring sessions across projects. | Sessions, chat, files, changes, validation and available Git actions. | Some features and harnesses are experimental or limited, and capabilities vary by configuration. |
A simple rule works well: stay in Chat when you are actively shaping one implementation; move to Agents when you are triaging several sessions, reviewing outcomes or checking work in different repositories.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Use the board for multi-project work without losing control
Give every session a testable outcome
Sessions are easier to monitor when each request names one outcome and its verification. “Investigate the login issue” is difficult to triage. “Reproduce the expired-session redirect, identify the responsible module, and add a regression test that demonstrates the corrected redirect” gives the agent and the reviewer a shared finish line.
Keep plans and implementation decisions together
Ask the agent to put important assumptions, rejected alternatives and verification results in the session conversation. When you return to the board later, the record explains why a file changed and which evidence supports completion.
Separate approval from delegation
Approving a plan is a design decision. Allowing an automated session to edit files, run commands, commit or create a pull request is an operational decision. Make each one consciously, especially for production branches or repositories containing secrets.
Use configured workspace actions carefully
The Agents window can support configured workspace tasks, terminal commands, commit actions and pull-request creation or follow-up options. Treat these as workflow capabilities, not as automatic quality checks. Before enabling or accepting one, confirm the target branch, command, credentials and expected review path.
Prerequisites and current limitations
Session and sign-in options
The documented Agents window supports GitHub Copilot, Claude API or BYOK configurations, Codex with an existing ChatGPT sign-in and other BYOK setups. Some sign-in-free paths are experimental. Your organization’s policies, installed extensions and account entitlements can change which choices appear.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallThe Plan agent is not in the agents dropdown
The agents dropdown currently does not list the Plan agent. Start planning with /plan in a Copilot or Claude session; in Copilot, a request for a plan may also trigger planning automatically. Do not search the dropdown for a separate Plan entry if it is absent.
Browser access
The documented browser entry point is available at https://insiders.vscode.dev/agents, and the documented flow requires GitHub sign-in. If your organization blocks the browser experience or the required sign-in, use the desktop Agents window instead.
Preview behavior
Automations are marked preview in VS Code documentation. Preview features can change labels, permissions, supported harnesses or behavior, so verify the controls shown by your installed version before relying on an automation in a repeatable team process.
Common failure modes and recovery steps
The plan is too broad
Symptom: the plan touches many unrelated modules or proposes a rewrite. Recovery: restate the smallest user-visible outcome, identify files that must remain unchanged and ask for a revised plan with explicit out-of-scope items.
Recommended Free Tools
The agent chose the wrong project or branch
Symptom: the Files view or Changes view shows an unexpected workspace or branch. Recovery: stop before implementation, open the intended workspace, and start or redirect the session there. Do not rely on a summary to compensate for an incorrect workspace.
Verification is vague or missing
Symptom: the plan says “test the change” without naming a command or scenario, or the agent reports completion without output. Recovery: specify the exact test, build, lint, task or manual flow required and ask for its result and limitations.
A test fails for an unrelated reason
Symptom: a command fails because a service, secret or runtime is unavailable. Recovery: ask the agent to separate environment failure from code failure, record the missing prerequisite and run the strongest available local check. Do not label the task verified until the relevant behavior has evidence.
The diff exceeds the approved scope
Symptom: Changes shows files or edits not present in the plan. Recovery: request an explanation, revert unrelated edits where appropriate and update the plan before accepting more work.
An automated action is surprising
Symptom: a terminal command, commit or pull-request action would affect a branch or system you did not intend to change. Recovery: decline the tool action, confirm the target and permissions, then rerun only the necessary step. Plan approval does not pre-approve every tool action.
A repeatable checklist
- Open the intended VS Code workspace and Chat session.
- Choose the applicable session target and enter
/plan, or select Plan from the agent-mode picker. - Describe the outcome, constraints and verification criteria.
- Answer clarifying questions and ask for trade-offs where choices matter.
- Review scope, affected files, dependencies, decisions and verification steps.
- Use overall or inline feedback to correct the plan, then reread the revision.
- Choose Implement Plan, Approve Plan Only, Implement with Autopilot or Reject deliberately.
- Open the Agents window when you need to monitor sessions across projects.
- Inspect Chat, Files and Changes for the selected session.
- Ask the agent to run relevant tests or exercise the application.
- Review command output, failures, limitations and the actual diff.
- Only then accept a commit, pull request or other follow-up action.
What “done” should mean
A coding task is ready for handoff when the implemented files match the approved scope, the requested behavior is demonstrated by relevant tests or a documented manual check, failures and environmental limitations are explicit, and the diff has been reviewed. The Agents window helps you see that evidence across sessions; it does not replace technical review.
For a complex change, follow the documented sequence: plan, implement, verify, then review. Keeping those stages distinct makes it easier to stop an unsafe edit, find a missing requirement and explain why a session is complete.
Quick Recap
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.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →
