Copilot Agent Manager

See exactly what your Copilot agents — and every subagent they spawn — are doing, live, without leaving VS Code

100% local · read-only · zero config

Copilot's chat window shows you the final answer, not the swarm behind it. Copilot Agent Manager turns the session events Copilot already writes into a calm, live dashboard — the full agent tree, what each agent is doing right now, what it changed, what it cost, and every session you've ever run — searchable in a keystroke.

Why you'll want it

• 👀 Never lose track of a subagent again. A live tree shows the main agent → every subagent it spawns, with status, current tool, and progress.
• 🧭 Understand a run at a glance instead of scrolling a chat transcript.
• 🔎 Find any past session by what was said in it — across every project — and resume it in one click.
• 💎 Native and safe. Pure VS Code styling, read-only, all on your machine. It never touches your running sessions.

✨ Features

🌳 Live agent & subagent tree

Your whole run, structured. The main agent with the subagents it spawned nested underneath — names and tasks pulled straight from Copilot's subagent.started events. Spinners for running, badges for done / error / waiting, token counts, and progress only when Copilot actually reports todos (never a fake bar).

⚡ "Right now" — the current tool call

For every running agent, see the exact tool in flight (edit src/auth/session.ts, bash npm test, grep …) plus a live streaming activity log that updates as it works.

💬 VS Code chat sessions, too

Editor chats don't live in the CLI session tree. Copilot Agent Manager reads both places VS Code keeps them: its native chat session files (workspaceStorage/*/chatSessions/*.jsonl, replayed from their incremental log so the latest turns and tool calls show up immediately) and Copilot's session-store.db (read directly, including turns still sitting in its write-ahead log). Sessions present in both are merged by id, so you get the full transcript alongside Copilot's summary and branch. They expose their turns, tool calls, and the files they referenced, but no subagents, so Continue stays hidden for them.

🔍 Search every session (full-text)

Hit the search icon and type a keyword. The extension scans the full text of every Copilot session's events with ripgrep — sub-second — and returns matching sessions (subagent hits folded in), each with a highlighted snippet. Then Continue or Read it in one click.

▶️ Continue any session

Pick up right where a session left off: Continue opens a terminal in the session's project and runs copilot --resume=<session-id>.

📂 Session-wide changes

One combined, PR-style view of every file the whole run touched — across the main agent and all subagents — with +/− counts and one-click diffs of exactly what changed.

📖 Readable conversations (not raw JSONL)

Open any agent's conversation as a clean, formatted Markdown view — prompts, replies, and tool calls — instead of squinting at events.jsonl.

🖥️ Sidebar or full-screen editor

Live in the activity bar, or pop the whole dashboard out into a full-screen editor tab — both stay in sync.

🎛️ Three layouts, one click

Switch between Tree (hierarchy + inline detail), Cards (one live card per agent), and Split (master list + Activity / Files / Conversation tabs).

🔔 Status bar + finish notifications

A glanceable status-bar chip (✦ 2 running · 1.2M tok) even when the panel is closed, plus a toast when an agent finishes or errors.

🚀 Getting started

1. Install Copilot Agent Manager from the Marketplace (or code --install-extension copilot-agent-manager-x.y.z.vsix).
2. Click the Copilot Agent Manager icon in the activity bar.
3. Run Copilot as usual — agent mode in VS Code or the Copilot CLI, ideally a task that spawns subagents — and watch it light up. ✨

🔎 Seeing nothing? First open is scoped to your current workspace. Click the scope toggle in the header (or run “Copilot Agents: Toggle Scope”) to show all projects.

No setup, no API keys, no sign-in. It reads the session events Copilot already writes under ~/.copilot/session-state, plus VS Code's own editor chat sessions (turn that off with includeChatStore).

⚙️ Settings

All under copilotAgentManager.*:

🛡️ Private & safe by design

• Read-only. It observes the session events Copilot already writes and opens VS Code's chat database read-only — it never sends input to, pauses, or interferes with your agents.
• 100% local. No network calls, no telemetry, no accounts. Your conversations never leave your machine.
• Fast on huge runs. Multi-megabyte session files are tail-read incrementally, never loaded whole in the hot path.

~/.copilot/session-state/<sessionId>/
  events.jsonl       ← every session event: prompts, messages, tool calls, subagents
  workspace.yaml     ← cwd, git branch, session name
  session.db         ← todos / inbox

…/User/workspaceStorage/<hash>/chatSessions/<sessionId>.jsonl
  snapshot line  ← session id, title, creation date, requests
  set / append   ← later turns, response parts and tool invocations

…/User/globalStorage/github.copilot-chat/session-store.db
  sessions       ← id, cwd, repository, branch, summary, agent name, timestamps
  turns          ← one row per exchange (user message + assistant response)
  session_files  ← files the session touched, with the tool that touched them

❓ FAQ

Does it work without any setup? Yes — if Copilot writes sessions to ~/.copilot/session-state, it just works.

Which Copilot surfaces are covered? Copilot CLI and VS Code's Copilot CLI runtime (full events, subagents, diffs, resume), plus VS Code's own editor chat sessions from its native chatSessions files and Copilot's session-store.db (turns, tool calls and referenced files; no subagents). Set copilotAgentManager.includeChatStore to false to ignore both.

Can it control or stop my agents? No, by design. It is a read-only monitor; Continue launches a fresh resume via copilot --resume, it never injects into a live run.

Does it send my data anywhere? Never. Everything stays on your machine.

📄 License

MIT © Damilola Onadeinde