FleetView

A read-only VS Code dashboard for a Claude Code agent fleet. FleetView reads the session and transcript files Claude Code already writes on your machine, so you can see every live main agent, every subagent it launched, what each one is writing right now, and what the main agent did with the result. GitHub Copilot Chat sessions are picked up the same way on Windows.

FleetView never launches, steers or edits an agent — it watches.

The Fleet view

Lives in its own Activity Bar container (FleetView → Fleet) and reveals itself when VS Code finishes starting.

Sessions are group rows: session name, workspace folder, and ▶ running ✓ done Σ total counts. Under each session, one entry per subagent, drawn as two lines:

• Task line — the agent's short id (the handle you paste back to the main agent), the task description, a model badge (Opus, Sonnet, … turning orange when the model changed mid-run, exact model ids on hover), 🌿 when the agent has its own git worktree, the agent type when it is not general-purpose, and recovered when the entry was rebuilt from disk rather than seen live.
• Metadata line — Status (colored dot plus running / done / failed / killed / stopped), Started, Duration, Last touch. Every column sorts, resizes and reorders; sessions collapse; the layout survives a reload.

Toolbar: a Running / All / Ended filter · Only me, which keeps the list to sessions whose working directory matches this window's workspace folders · Empty agents, which keeps sessions that have nothing matching the current filter · a sort selector (Newest, Oldest, Status, Longest, Recently active, Task A–Z) · collapse-all · and a refresh button in the view title bar.

Agent detail — Output · Diff · Harvest

Click a subagent to open its panel.

• Output — tails that agent's own JSONL transcript. The last 256 KB is read up front, then appends stream in live; assistant prose, 🔧 tool calls and ↩︎ tool results are rendered apart from each other and non-JSON lines fall through raw. It follows the tail and pauses the moment you scroll up.
• Diff — git diff of the agent's worktree against merge-base(HEAD, <default branch>), so committed and still-uncommitted work both appear, untracked files included. It re-diffs every 6 seconds while the agent is alive. Hover any changed line and click 💬 to compose a review comment pre-filled with the agent id and file:line, then Copy for main agent to hand it over before the work is merged. An agent that never edits files, and a worktree that has already been merged and pruned, each say so rather than showing an empty diff. Rendering is capped at 60 files and 500 lines per file.
• Harvest — what the subagent handed back and its terminal status, the main agent's next turn, and the tool calls it queued off the back of it — reconstructed from the parent session transcript, never the subagent's own.

Fleet map

⌗ map on a session row opens the attribution table for that session: Subagent, Agent ID, Type, Model, Worktree branch, Files, Δ lines, State — worktree agents first, then by lines changed, so parallel agents never blur together. Click a row to open that agent, click an id to copy it. It refreshes every 15 seconds, and git work is capped at 40 agents per pass so opening the map cannot stall the extension host.

Status bar and notifications

A 🚢 N agents item sits in the status bar at all times, mirrors the Only me scope so the count and the visible rows never disagree, and opens the Fleet view when clicked.

When a subagent settles, a toast reports its outcome and duration with View harvest and View output buttons that open the panel on that tab. Toasts are raised only for sessions belonging to this window, so a busy machine does not spam you from unrelated workspaces.

Commands

What it reads

Requires VS Code 1.86 or newer. There is nothing to configure.

Privacy

Everything stays on the machine — FleetView makes no network calls of any kind. At activation it writes a snapshot of the sessions it found to fleetview-diag.txt in your system temp directory.