@esso0428/pi-subagents 0.16.0 → 0.17.1

package/CHANGELOG.md

@@ -7,6 +7,16 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.17.1] - 2026-09-15
+
+### Fixed
+- **The Agents panel now initializes during TUI session startup after restoring persisted agent records**, so openable completed-agent history is visible immediately when a session starts or resumes instead of waiting for the next tool execution.
+
+## [0.17.0] - 2026-09-15
+
+### Changed
+- **The above-editor Agents panel is now the single agent navigator.** It replaces the duplicate FleetView roster, uses focus-gated `↑`/`↓` navigation with `Enter` to open live or read-only history viewers, and removes FleetView's duplicate timer and deferred redraw path.
+
 ## [0.16.0] - 2026-09-15
 
 > **⚠️ Breaking: pi 0.80.8 is now the minimum supported host version.** The package peer dependencies require `@earendil-works/pi-ai`, `@earendil-works/pi-coding-agent`, and `@earendil-works/pi-tui` `>=0.80.8`; upgrade pi before installing this release. This floor matches the `ModelRuntime` API used by the runner and avoids installing a release against the pre-0.80.8 session-creation contract.

package/README.md

@@ -14,8 +14,7 @@ https://github.com/user-attachments/assets/8685261b-9338-4fea-8dfe-1c590d5df543
 
 - **Claude Code look & feel** — same tool names, calling conventions, and UI patterns (`Agent`, `get_subagent_result`, `steer_subagent`) — feels native
 - **Parallel background agents** — spawn multiple agents that run concurrently with automatic queuing (configurable concurrency limit, default 4) and smart group join (consolidated notifications)
-- **Live widget UI** — persistent above-editor widget with animated spinners, live tool activity, token counts, and colored status icons. Configurable via `/agents → Settings → Widget`: `all` (every agent), `background` (default — hides foreground runs, which already render inline as the `Agent` tool result), or `off`
-- **FleetView** — Claude Code-style navigable list of `main` + every running subagent rendered below the editor (earliest-launched first). Press `↓` (or `←`) at an empty prompt to jump in, `↑`/`↓`/wheel to move the selection, `Enter` to open the selected agent's live, auto-updating conversation, `Esc` to return. The selected row stays visible while the bounded list window moves through the full roster. Finished agents linger briefly before dropping out, and a viewer stays open through completion so you can read the final output. Toggle via `/agents → Settings → Fleet view`
+- **Agents panel UI** — persistent above-editor widget with animated spinners, live tool activity, token counts, colored status icons, and one focus-gated navigator. It shows every agent by default; press `↓` at an empty prompt to activate the panel, `↑`/`↓` to select, `Enter` to open a live or read-only history viewer, and `Esc` to return. Configure via `/agents → Settings → Widget`: `all`, `background`, or `off`
 - **Conversation viewer** — select any agent in `/agents` to open a live-scrolling overlay of its full conversation (auto-follows new content, scroll up to pause). Steer a running agent inline by pressing `Enter` to open a composer, typing, then `Enter` to send (`Esc` or an empty submit returns) — the message appears as a user message and redirects the agent after its current tool. Stop a still-running agent by pressing `x` (then `x` again to confirm) — both work for background agents too
 - **Custom agent types** — define agents in `.pi/agents/<name>.md` or `.agents/agents/<name>.md` (project) or globally, with YAML frontmatter: custom system prompts, model selection, thinking levels, tool restrictions
 - **Mid-run steering** — inject messages into running agents to redirect their work without restarting
@@ -95,7 +94,7 @@ Restrictions:
 
 ## UI
 
-The extension renders a persistent widget above the editor showing active agents. By default it shows background runs only (`widgetMode: background`) — foreground agents already render inline as the `Agent` tool result, so the widget would otherwise double-render them. Switch to `all` (every agent) or `off` (hide the widget) via `/agents → Settings → Widget`:
+The extension renders a persistent Agents widget above the editor. It initializes during TUI `session_start`, after the session branch restores completed-agent history, so openable agents are visible immediately when a session starts or resumes. By default it shows all agents (`widgetMode: all`), including foreground and background runs. Switch to `background` or `off` via `/agents → Settings → Widget`:
 
 ```
 ● Agents
@@ -112,20 +111,19 @@ The token field is annotated with two optional signals inside parens:
 - **`NN%`** — context-window utilization (color-coded: <70% dim, 70–85% warning, ≥85% error). Omitted when the model has no declared `contextWindow`, or briefly right after compaction.
 - **`⇊N`** — number of times the session has compacted, when > 0. Stays dim; the percent's color carries urgency.
 
-### FleetView
+### Agents panel navigator
 
-While subagents are running, a Claude Code-style navigable list renders **below** the editor:
+While agents are running or have openable history, the single Agents panel renders **above** the editor:
 
 ```
-  esc to interrupt · ← for agents · ↓ to manage
-
-  ● main
-  ○ general-purpose  Sleep then report 1                                11s · ↓ 13.1k tokens
-  ○ general-purpose  Sleep then report 2                                11s · ↓ 13.1k tokens
-                                                                                   ↓ 3 more
+● Agents  ↑↓ select · enter view · esc back
+├─ ● Agent  Refactor auth module · ↻5≤30 · 5 tool uses · 12.3s
+│    ⎿  editing 2 files…
+├─ ○ Explore  Find auth files · ↻3 · 4.1s
+└─ +1 more (↓ 1 more; 1 finished)
 ```
 
-The list is ordered earliest-launched first, and only shows agents you can actually open (pending/queued agents with no session yet appear once they start). At an **empty prompt**, press `↓` (or `←`) to move focus from the prompt into the list — the selected row is marked `●`, the rest `○`. `↑`/`↓` move the selection, `Enter` opens the selected agent's live conversation overlay (it auto-updates as the agent works), and `Esc` (or `↑` above `main`) returns to the prompt. Selecting `main` returns to the normal view. Inside the overlay, press `Enter` to steer the running agent — type a message and `Enter` to send it (`Esc` or an empty submit returns), and it redirects the agent the same way the `steer_subagent` tool does. A viewer stays open when its agent finishes so you can read the final output, and finished agents linger in the list for a few seconds before dropping out. Typing anything at a non-empty prompt behaves normally — the list only captures arrow keys when the prompt is empty. Disable it entirely via `/agents → Settings → Fleet view`.
+At an **empty prompt**, press `↓` to activate the panel. The selected row is marked `●`; `↑`/`↓` move through the full bounded roster, `Enter` opens a running agent live or a terminal agent's read-only history. A queued agent remains selectable, but has no live viewer until it starts; `Esc` returns to the prompt. `↑` while the first row is selected also returns to the prompt. Pressing `↑` at an inactive prompt, typing in a non-empty prompt, or using `j`/`k`/`←` behaves normally; the panel only captures the focus-gated arrow sequence described above. Disable it entirely via `/agents → Settings → Widget → off`.
 
 Individual agent results render Claude Code-style in the conversation:
 

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAIH,OAAO,EAAc,KAAK,YAAY,EAA0F,MAAM,iCAAiC,CAAC;AACxK,OAAO,EAAE,SAAS,EAAiE,MAAM,wBAAwB,CAAC;AAmBlH,OAAO,EAcL,KAAK,KAAK,EAEX,MAAM,sBAAsB,CAAC;AA8C9B,wBAAgB,wBAAwB,CACtC,KAAK,EAAE,MAAM,EACb,SAAS,EAAE,MAAM,EACjB,QAAQ,EAAE,MAAM,EAChB,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GACvB,SAAS,CAKX;AAkKD,MAAM,CAAC,OAAO,WAAW,EAAE,EAAE,YAAY,QAqtExC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAIH,OAAO,EAAc,KAAK,YAAY,EAA0F,MAAM,iCAAiC,CAAC;AACxK,OAAO,EAAE,SAAS,EAAiE,MAAM,wBAAwB,CAAC;AAmBlH,OAAO,EAcL,KAAK,KAAK,EAEX,MAAM,sBAAsB,CAAC;AA4C9B,wBAAgB,wBAAwB,CACtC,KAAK,EAAE,MAAM,EACb,SAAS,EAAE,MAAM,EACjB,QAAQ,EAAE,MAAM,EAChB,KAAK,EAAE,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GACvB,SAAS,CAKX;AAkKD,MAAM,CAAC,OAAO,WAAW,EAAE,EAAE,YAAY,QAuqExC"}

package/dist/index.js

@@ -31,8 +31,6 @@ import { resolveStorePath, ScheduleStore } from "./schedule-store.js";
 import { applyAndEmitLoaded, saveAndEmitChanged } from "./settings.js";
 import { getStatusNote } from "./status-note.js";
 import { AgentWidget, buildInvocationTags, describeActivity, fgPreservingNestedStyles, formatDuration, formatMs, formatTokens, formatTurns, getDisplayName, getPromptModeLabel, SPINNER, } from "./ui/agent-widget.js";
-import { createDeferredUiRefresh } from "./ui/deferred-ui-refresh.js";
-import { FleetList } from "./ui/fleet-list.js";
 import { showSchedulesMenu } from "./ui/schedule-menu.js";
 import { addUsage, getLifetimeTotal, getSessionContextPercent } from "./usage.js";
 // ---- Shared helpers ----
@@ -320,17 +318,14 @@ export default function (pi) {
     function sendIndividualNudge(record) {
         agentActivity.delete(record.id);
         widget.markFinished(record.id);
-        fleet.onAgentFinished(record.id);
         scheduleNudge(record.id, () => emitIndividualNudge(record));
         widget.update();
-        scheduleUiRefresh();
     }
     // ---- Group join manager ----
     const groupJoin = new GroupJoinManager((records, partial) => {
         for (const r of records) {
             agentActivity.delete(r.id);
             widget.markFinished(r.id);
-            fleet.onAgentFinished(r.id);
         }
         const groupKey = `group:${records.map(r => r.id).join(",")}`;
         scheduleNudge(groupKey, () => {
@@ -338,7 +333,6 @@ export default function (pi) {
             const unconsumed = records.filter(r => !r.resultConsumed);
             if (unconsumed.length === 0) {
                 widget.update();
-                scheduleUiRefresh();
                 return;
             }
             const notifications = unconsumed.map(r => formatTaskNotification(r, 300)).join('\n\n');
@@ -358,7 +352,6 @@ export default function (pi) {
             }, { deliverAs: "followUp", triggerTurn: true });
         });
         widget.update();
-        scheduleUiRefresh();
     }, 30_000);
     /** Helper: build event data for lifecycle events from an AgentRecord. */
     function buildEventData(record) {
@@ -409,16 +402,13 @@ export default function (pi) {
         if (record.resultConsumed) {
             agentActivity.delete(record.id);
             widget.markFinished(record.id);
-            fleet.onAgentFinished(record.id);
             widget.update();
-            scheduleUiRefresh();
             return;
         }
         // If this agent is pending batch finalization (debounce window still open),
         // don't send an individual nudge — finalizeBatch will pick it up retroactively.
         if (currentBatchAgents.some(a => a.id === record.id)) {
             widget.update();
-            scheduleUiRefresh();
             return;
         }
         const result = groupJoin.onAgentComplete(record);
@@ -428,7 +418,6 @@ export default function (pi) {
         // 'held' → do nothing, group will fire later
         // 'delivered' → group callback already fired
         widget.update();
-        scheduleUiRefresh();
     }, undefined, (record) => {
         // Emit started event when agent transitions to running (including from queue)
         pi.events.emit("subagents:started", {
@@ -438,9 +427,6 @@ export default function (pi) {
         });
         widget.ensureTimer();
         widget.update();
-        fleet.ensureTimer();
-        fleet.update();
-        scheduleUiRefresh();
     }, (record, info) => {
         // Emit compacted event when agent's session compacts (preserves count on record).
         pi.events.emit("subagents:compacted", {
@@ -513,6 +499,12 @@ export default function (pi) {
         manager.restoreCompleted(branch
             .filter((entry) => entry?.type === "custom" && entry?.customType === "subagents:record")
             .map((entry) => entry.data));
+        // Attach the panel during TUI startup, after restored records are present,
+        // so terminal agents from the session branch are immediately visible.
+        if (ctx.mode === "tui") {
+            widget.setUICtx(ctx.ui);
+            widget.update();
+        }
         // Guard mirrors the `!scheduler.isActive()` pattern below: session_start
         // fires once per activation, but a double-bind must not leak listeners.
         if (!rpcHandle) {
@@ -531,7 +523,6 @@ export default function (pi) {
             startScheduler(ctx);
     });
     pi.on("session_before_switch", () => {
-        deferredUiRefresh.cancel();
         manager.clearCompleted(true);
         scheduler.stop();
     });
@@ -553,45 +544,23 @@ export default function (pi) {
         for (const timer of pendingNudges.values())
             clearTimeout(timer);
         pendingNudges.clear();
-        deferredUiRefresh.dispose();
-        fleet.dispose();
+        widget.dispose();
         manager.dispose();
     });
-    // Live widget: show running agents above editor.
-    // widgetMode (default "background") selects what the widget shows: "all" =
-    // every agent; "background" = hide foreground (they already render inline as
-    // the Agent tool result, so showing them here too is a duplicate, #118), keep
-    // everything else; "off" = hide the widget entirely. Read live at render time.
-    let widgetMode = "background";
+    // Live widget: show all agents above the editor. Read live at render time.
+    let widgetMode = "all";
     function getWidgetMode() { return widgetMode; }
-    const widget = new AgentWidget(manager, agentActivity, getWidgetMode);
+    const widget = new AgentWidget(manager, agentActivity, getWidgetMode, {
+        canOpenHistory: (record) => canOpenAgentHistory(record, currentCtx?.cwd),
+        onOpen: (record, mode) => {
+            const ctx = currentCtx;
+            if (ctx)
+                void viewAgentConversation(ctx, record, mode);
+        },
+    });
     function setWidgetMode(m) {
-        const changed = widgetMode !== m;
         widgetMode = m;
         widget.update();
-        if (changed)
-            scheduleUiRefresh();
-    }
-    // Claude Code-style FleetView: navigable list of main + subagents below the editor.
-    const fleet = new FleetList(manager, agentActivity, () => currentCtx?.cwd, pi, () => currentCtx);
-    // One render-only scheduler is shared by both widgets. The callback resolves
-    // the live TUI target at fire time because session switches invalidate the
-    // previous widget context; both widgets use the same TUI, so one successful
-    // request is sufficient.
-    const deferredUiRefresh = createDeferredUiRefresh(() => {
-        if (widget.requestUiRefresh(true))
-            return;
-        fleet.requestUiRefresh(true);
-    });
-    const scheduleUiRefresh = () => deferredUiRefresh.schedule();
-    let fleetViewEnabled = true;
-    function isFleetViewEnabled() { return fleetViewEnabled; }
-    function setFleetViewEnabled(b) {
-        const changed = fleetViewEnabled !== b;
-        fleetViewEnabled = b;
-        fleet.setEnabled(b);
-        if (changed)
-            scheduleUiRefresh();
     }
     // Project/global default for writing the subagent .output transcript. A custom
     // agent's `output_transcript` frontmatter overrides this per spawn; when the
@@ -684,10 +653,7 @@ export default function (pi) {
     }
     // Grab UI context from first tool execution + clear lingering widget on new turn
     pi.on("tool_execution_start", async (_event, ctx) => {
-        const widgetContextChanged = widget.setUICtx(ctx.ui);
-        const fleetContextChanged = fleet.setUICtx(ctx.ui);
-        if (widgetContextChanged || fleetContextChanged)
-            scheduleUiRefresh();
+        widget.setUICtx(ctx.ui);
         widget.onTurnStart();
     });
     /** Format an agent's tool scope: "*" when it has all built-ins, else a comma-separated list. */
@@ -738,7 +704,6 @@ export default function (pi) {
         setScopeModels: setScopeModelsEnabled,
         setDisableDefaultAgents: setDisableDefaultAgents,
         setToolDescriptionMode: setToolDescriptionMode,
-        setFleetView: setFleetViewEnabled,
         setWidgetMode: setWidgetMode,
         setOutputTranscript: setOutputTranscript,
     }, (event, payload) => pi.events.emit(event, payload));
@@ -1217,9 +1182,6 @@ Terse command-style prompts produce shallow, generic work.
                 agentActivity.set(id, bgState);
                 widget.ensureTimer();
                 widget.update();
-                fleet.ensureTimer();
-                fleet.update();
-                scheduleUiRefresh();
                 // Emit created event
                 pi.events.emit("subagents:created", {
                     id,
@@ -1272,9 +1234,6 @@ Terse command-style prompts produce shallow, generic work.
                         agentActivity.set(a.id, fgState);
                         widget.ensureTimer();
                         widget.update();
-                        fleet.ensureTimer();
-                        fleet.update();
-                        scheduleUiRefresh();
                         break;
                     }
                 }
@@ -1322,7 +1281,6 @@ Terse command-style prompts produce shallow, generic work.
             if (fgId) {
                 agentActivity.delete(fgId);
                 widget.markFinished(fgId);
-                fleet.onAgentFinished(fgId);
             }
             // Get final token count
             const tokenText = formatLifetimeTokens(fgState);
@@ -2091,7 +2049,6 @@ ${systemPrompt}
             scopeModels: isScopeModelsEnabled(),
             disableDefaultAgents: isDefaultsDisabled(),
             toolDescriptionMode: getToolDescriptionMode(),
-            fleetView: isFleetViewEnabled(),
             widgetMode: getWidgetMode(),
             outputTranscript: getOutputTranscriptDefault(),
         };
@@ -2159,13 +2116,6 @@ ${systemPrompt}
                     currentValue: getOutputTranscriptDefault() ? "on" : "off",
                     values: ["on", "off"],
                 },
-                {
-                    id: "fleetView",
-                    label: "Fleet view",
-                    description: "Claude Code-style main+subagents list below the editor (↓/← to navigate, Enter to view)",
-                    currentValue: isFleetViewEnabled() ? "on" : "off",
-                    values: ["on", "off"],
-                },
                 {
                     id: "widgetMode",
                     label: "Widget",
@@ -2243,11 +2193,6 @@ ${systemPrompt}
                 setToolDescriptionMode(value);
                 notifyApplied(ctx, `Tool description set to ${value}. Takes effect on next pi session.`);
             }
-            else if (id === "fleetView") {
-                const enabled = value === "on";
-                setFleetViewEnabled(enabled);
-                notifyApplied(ctx, `Fleet view ${enabled ? "enabled" : "disabled"}`);
-            }
             else if (id === "widgetMode") {
                 setWidgetMode(value);
                 notifyApplied(ctx, `Widget set to ${value}`);