@pi-unipi/unipi 3.0.0-alpha.22 → 3.0.0-alpha.23

package/CHANGELOG.md

@@ -6,6 +6,27 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 
 ## [Unreleased]
 
+## [3.0.0-alpha.23] — 2026-10-04
+
+### Added
+- `core`: the `/unipi:settings` hub explains the selected setting in a fixed 4-row info panel — wrapped description, the highlighted (or stored) enum option's meaning, and a meta line (`default … · range … · scope`). Enum and multiselect options carry their own descriptions, shown while the option list is open; number fields take `unit` (`6000 ms`, `80%`); option descriptions join the search haystack. (UNI-50)
+- guard test (`tests/settings-metadata.test.ts`): every settings field and option must be described (labels ≤ 24 chars, sentences end with a period, units live in `unit`, option fragments don't) across all 23 namespaces.
+
+### Changed
+- `web-api`: web tools that cannot work are no longer exposed to the agent — `web_search`/`web_llm_summarize` drop out of the active tools when no enabled provider can serve them (re-checked every prompt), `multi_web_content_read` always stays (smart-fetch is local). The phantom `llm-summarize` provider is gone and `web_llm_summarize` lost its useless `source` parameter. Tool failures now throw instead of returning ignored `isError` results. (UNI-72)
+- settings: shorter labels (explanations moved into descriptions), and pruned knobs —
+  - `background-tasks`: removed `defaultTimeoutSeconds` (nothing read it).
+  - `compactor`: the nine `sections.*` booleans became one **Summary sections** multiselect (stored booleans still work); `cooldownMs`/`repeatMinGrowthTokens` fixed at their defaults; `debug` moved to `UNIPI_COMPACTOR_DEBUG=1`.
+  - `memory`: `recallAtStart` + `wakeUp` merged into one **Recall at start** enum (`off`/`reminder`/`wake-up`); stored booleans migrate per scope.
+  - `permission`: removed `jevJudge` — auto mode always consults jev.
+  - `footer`: classic-only options (`preset`, `separator`, `zoneSeparator`, `full labels`, `color mode`) moved under an advanced **Classic footer** section; preset list now offers the real presets (`compact`, `full`, `ascii` instead of the dead `dense`/`devops`/`zen`).
+  - `image`: custom endpoint fields moved into per-tool **Custom endpoint…** pages (summary shows `off` or the host).
+  - `web-api`: Smart fetch section is advanced; `os`/`batchConcurrency` removed from the schema (engine defaults remain).
+  - `kanboard`: reorganized into Agent / Housekeeping / **Board server** (advanced, plus open/stop/rotate-token actions); dead `blocking` setting deleted.
+
+### Docs
+- README: dropped the "Why we have it" column from the Proof table.
+
 ## [3.0.0-alpha.22] — 2026-10-03
 
 ### Changed

package/README.md

@@ -68,19 +68,18 @@ runs commands. UniPi adds the parts that a long session needs.
 ## Proof
 
 Each feature below exists because published results support it. The numbers
-come from the linked sources. They are not UniPi benchmarks. The "why" column
-is our opinion.
-
-| Feature | Benchmark | Why we have it | Source |
-|---|---|---|---|
-| **Memory** | 96.6% recall@5 on LongMemEval, with raw semantic search and 0 API calls. | Chat assistants lose about 30% accuracy when facts span many sessions. We want the agent to find last week's decision without a model call. MemPalace does this on your machine. | [MemPalace](https://github.com/MemPalace/mempalace) · [LongMemEval](https://arxiv.org/abs/2410.10813) |
-| **Fusion** | Frontier-level FrontierCode score at up to 60% lower cost. | Most turns in a coding session are routine edits and commands. We think a frontier model should plan and review, and a lower-cost model should do the typing. | [Devin Fusion](https://cognition.com/blog/devin-fusion) |
-| **Fusion with two frontier models** | 69.0% on DRACO for two frontier models together. The best single model scored 65.3%. | Different models make different mistakes. A second strong model finds errors that one model misses. Use two frontier models when quality matters more than cost. | [OpenRouter Fusion](https://openrouter.ai/blog/announcements/fusion-beats-frontier/) · [Mixture-of-Agents](https://arxiv.org/abs/2406.04692) |
-| **Goal** | HumanEval pass@1 went from 80% to 91% with a loop on test feedback. | An agent that grades its own work stops too early or claims success. Self-written checks can hide real failures. Our goal loop continues until a separate verifier agrees, inside turn and stall budgets. | [Reflexion](https://arxiv.org/abs/2303.11366) · [Self-authored verification](https://arxiv.org/abs/2607.24300) |
-| **Ralph** | Six repositories overnight for $297 in API cost (a reported hackathon result). | Long tasks do not fit in one context window. A checklist file keeps the progress on disk. Each iteration works the next items and checks them off. A verifier reads the file before the loop ends. We think a simple loop is the right tool for checklist work. | [Ralph loop](https://ghuntley.com/ralph/) · [YC hackathon report](https://paddo.dev/blog/ralph-wiggum-autonomous-loops/) |
-| **Swarm** | 90.2% higher than a single agent on Anthropic's research eval. | Independent items do not need to wait for each other. Each worker gets a clean context, and the lead reads only the summaries. A swarm uses more tokens, so we keep it for wide work. | [Anthropic multi-agent research](https://www.anthropic.com/engineering/multi-agent-research-system) |
-| **Graph** | Up to 3.7× lower latency, 6.7× lower cost and about 9% higher accuracy than ReAct. | Many tasks have steps that need earlier results. A graph runs the independent steps at the same time. Each step gets only the inputs it needs. | [LLMCompiler (ICML 2024)](https://arxiv.org/abs/2312.04511) |
-| **Prefix cache** | Up to 90% lower cost and 85% lower latency for long prompts. | Each agent turn sends the full conversation again. In our UniPi 2.4 logs, each cache miss sent about 62,700 tokens again at the full price. UniPi keeps the prefix byte-identical, so a turn pays full price only for the new tail. | [Anthropic prompt caching](https://claude.com/blog/prompt-caching) · [Modular handbook](https://handbook.modular.com/inference-optimization/prefix-caching/) · [our cache study](docs/deepseek-cache-rate-research.md) |
+come from the linked sources. They are not UniPi benchmarks.
+
+| Feature | Benchmark | Source |
+|---|---|---|
+| **Memory** | 96.6% recall@5 on LongMemEval, with raw semantic search and 0 API calls. | [MemPalace](https://github.com/MemPalace/mempalace) · [LongMemEval](https://arxiv.org/abs/2410.10813) |
+| **Fusion** | Frontier-level FrontierCode score at up to 60% lower cost. | [Devin Fusion](https://cognition.com/blog/devin-fusion) |
+| **Fusion with two frontier models** | 69.0% on DRACO for two frontier models together. The best single model scored 65.3%. | [OpenRouter Fusion](https://openrouter.ai/blog/announcements/fusion-beats-frontier/) · [Mixture-of-Agents](https://arxiv.org/abs/2406.04692) |
+| **Goal** | HumanEval pass@1 went from 80% to 91% with a loop on test feedback. | [Reflexion](https://arxiv.org/abs/2303.11366) · [Self-authored verification](https://arxiv.org/abs/2607.24300) |
+| **Ralph** | Six repositories overnight for $297 in API cost (a reported hackathon result). | [Ralph loop](https://ghuntley.com/ralph/) · [YC hackathon report](https://paddo.dev/blog/ralph-wiggum-autonomous-loops/) |
+| **Swarm** | 90.2% higher than a single agent on Anthropic's research eval. | [Anthropic multi-agent research](https://www.anthropic.com/engineering/multi-agent-research-system) |
+| **Graph** | Up to 3.7× lower latency, 6.7× lower cost and about 9% higher accuracy than ReAct. | [LLMCompiler (ICML 2024)](https://arxiv.org/abs/2312.04511) |
+| **Prefix cache** | Up to 90% lower cost and 85% lower latency for long prompts. | [Anthropic prompt caching](https://claude.com/blog/prompt-caching) · [Modular handbook](https://handbook.modular.com/inference-optimization/prefix-caching/) · [our cache study](docs/deepseek-cache-rate-research.md) |
 
 Long work matters more each year. METR measured that the length of software
 tasks that agents can finish doubles about every 7 months

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pi-unipi/unipi",
-  "version": "3.0.0-alpha.22",
+  "version": "3.0.0-alpha.23",
   "description": "All-in-one extension suite for Pi coding agent",
   "type": "module",
   "license": "MIT",
@@ -27,7 +27,7 @@
     "prepare": "node scripts/build-bundle.mjs --if-missing",
     "prepublishOnly": "npm run build",
     "publish:all": "npm publish --workspaces --access public",
-    "test": "npx tsx --test tests/*.test.js packages/memory/tests/*.test.ts && npm test --workspaces --if-present",
+    "test": "npx tsx --test tests/*.test.js tests/*.test.ts packages/memory/tests/*.test.ts && npm test --workspaces --if-present",
     "typecheck": "npx tsc --noEmit --skipLibCheck",
     "tui:gallery": "npx tsx scripts/tui-gallery/index.ts",
     "publish:kanboard": "node scripts/publish-kanboard-platforms.mjs"
@@ -82,26 +82,26 @@
     "typebox": "^1.1.38"
   },
   "dependencies": {
-    "@pi-unipi/ask-user": "3.0.0-alpha.22",
-    "@pi-unipi/background-tasks": "3.0.0-alpha.22",
-    "@pi-unipi/btw": "3.0.0-alpha.22",
-    "@pi-unipi/command-enchantment": "3.0.0-alpha.22",
-    "@pi-unipi/compactor": "3.0.0-alpha.22",
-    "@pi-unipi/core": "3.0.0-alpha.22",
-    "@pi-unipi/footer": "3.0.0-alpha.22",
-    "@pi-unipi/fusion": "3.0.0-alpha.22",
-    "@pi-unipi/info-screen": "3.0.0-alpha.22",
-    "@pi-unipi/input-shortcuts": "3.0.0-alpha.22",
-    "@pi-unipi/kanboard": "3.0.0-alpha.22",
-    "@pi-unipi/mcp": "3.0.0-alpha.22",
-    "@pi-unipi/memory": "3.0.0-alpha.22",
-    "@pi-unipi/notify": "3.0.0-alpha.22",
-    "@pi-unipi/skill-registry": "3.0.0-alpha.22",
-    "@pi-unipi/subagents": "3.0.0-alpha.22",
-    "@pi-unipi/updater": "3.0.0-alpha.22",
-    "@pi-unipi/utility": "3.0.0-alpha.22",
-    "@pi-unipi/web-api": "3.0.0-alpha.22",
-    "@pi-unipi/workflow": "3.0.0-alpha.22"
+    "@pi-unipi/ask-user": "3.0.0-alpha.23",
+    "@pi-unipi/background-tasks": "3.0.0-alpha.23",
+    "@pi-unipi/btw": "3.0.0-alpha.23",
+    "@pi-unipi/command-enchantment": "3.0.0-alpha.23",
+    "@pi-unipi/compactor": "3.0.0-alpha.23",
+    "@pi-unipi/core": "3.0.0-alpha.23",
+    "@pi-unipi/footer": "3.0.0-alpha.23",
+    "@pi-unipi/fusion": "3.0.0-alpha.23",
+    "@pi-unipi/info-screen": "3.0.0-alpha.23",
+    "@pi-unipi/input-shortcuts": "3.0.0-alpha.23",
+    "@pi-unipi/kanboard": "3.0.0-alpha.23",
+    "@pi-unipi/mcp": "3.0.0-alpha.23",
+    "@pi-unipi/memory": "3.0.0-alpha.23",
+    "@pi-unipi/notify": "3.0.0-alpha.23",
+    "@pi-unipi/skill-registry": "3.0.0-alpha.23",
+    "@pi-unipi/subagents": "3.0.0-alpha.23",
+    "@pi-unipi/updater": "3.0.0-alpha.23",
+    "@pi-unipi/utility": "3.0.0-alpha.23",
+    "@pi-unipi/web-api": "3.0.0-alpha.23",
+    "@pi-unipi/workflow": "3.0.0-alpha.23"
   },
   "devDependencies": {
     "@earendil-works/pi-agent-core": "^0.87.1",

package/packages/ask-user/config.ts

@@ -44,9 +44,9 @@ registerSettings({
     {
       title: "Tool",
       fields: [
-        { key: "enabled", type: "boolean", label: "Enable ask_user", description: "Let the agent stop and ask you multiple-choice questions" },
-        { key: "notifyOnAsk", type: "boolean", label: "Notify when asked", description: "Send a notification when the agent is waiting for your answer" },
-        { key: "maxQuestions", type: "number", label: "Questions per call", min: 1, max: 4, description: "How many questions one dialog may hold (applies to new sessions)" },
+        { key: "enabled", type: "boolean", label: "Enable ask_user", description: "Let the agent stop and ask you multiple-choice questions." },
+        { key: "notifyOnAsk", type: "boolean", label: "Notify when asked", description: "Send a notification when the agent is waiting for your answer." },
+        { key: "maxQuestions", type: "number", label: "Questions per call", min: 1, max: 4, description: "How many questions one dialog may hold. Applies to new sessions." },
       ],
     },
     {
@@ -58,24 +58,24 @@ registerSettings({
           type: "enum",
           label: "Esc",
           options: [
-            { value: "stop", label: "stop the agent's turn" },
-            { value: "send", label: "send what's answered (rest skipped)" },
+            { value: "stop", label: "stop", description: "stop the agent's turn" },
+            { value: "send", label: "send", description: "send what's answered, skip the rest" },
           ],
-          description: "What Esc does in the dialog",
+          description: "What Esc does in the dialog.",
         },
-        { key: "digitAdvance", type: "boolean", label: "Number keys move on", description: "In a single-choice question, picking with 1–9 also goes to the next question" },
+        { key: "digitAdvance", type: "boolean", label: "Number keys move on", description: "In a single-choice question, picking with 1–9 also advances." },
         {
           key: "other",
           type: "enum",
           label: "\"Other\" choice",
           options: [
-            { value: "agent", label: "agent decides (default on)" },
-            { value: "always", label: "always offer it" },
-            { value: "never", label: "never offer it" },
+            { value: "agent", label: "agent decides", description: "on unless the agent opts out" },
+            { value: "always", label: "always", description: "always offer the Other row" },
+            { value: "never", label: "never", description: "never offer the Other row" },
           ],
-          description: "The free-text \"Other (type your own)\" row",
+          description: "The free-text \"Other (type your own)\" row.",
         },
-        { key: "helpLine", type: "boolean", label: "\"Not ready\" line", description: "Show \"? Not ready to answer, help me out!\" under the dialog" },
+        { key: "helpLine", type: "boolean", label: "\"Not ready\" line", description: "Show \"? Not ready to answer, help me out!\" under the dialog." },
       ],
     },
   ],

package/packages/ask-user/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pi-unipi/ask-user",
-  "version": "3.0.0-alpha.22",
+  "version": "3.0.0-alpha.23",
   "description": "Structured user input tool for Pi coding agent — single-select, multi-select, freeform",
   "type": "module",
   "main": "index.ts",
@@ -40,7 +40,7 @@
     "access": "public"
   },
   "dependencies": {
-    "@pi-unipi/core": "3.0.0-alpha.22"
+    "@pi-unipi/core": "3.0.0-alpha.23"
   },
   "peerDependencies": {
     "@earendil-works/pi-coding-agent": "^0.87.1",

package/packages/autocomplete/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pi-unipi/command-enchantment",
-  "version": "3.0.0-alpha.22",
+  "version": "3.0.0-alpha.23",
   "description": "Enhanced TUI autocomplete for /unipi:* commands — colored, sorted, and grouped by package",
   "type": "module",
   "main": "src/index.ts",

package/packages/autocomplete/src/settings.ts

@@ -29,7 +29,7 @@ registerSettings({
           key: "autocompleteEnhanced",
           type: "boolean",
           label: "Enhanced autocomplete",
-          description: "Fuzzy /unipi:* command suggestions",
+          description: "Fuzzy /unipi:* command suggestions while you type.",
         },
       ],
     },

package/packages/background-tasks/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pi-unipi/background-tasks",
-  "version": "3.0.0-alpha.22",
+  "version": "3.0.0-alpha.23",
   "description": "Background tasks for UniPi — durable shell jobs and delegated background agents with a footer dock",
   "type": "module",
   "main": "src/index.ts",
@@ -10,7 +10,7 @@
   },
   "dependencies": {
     "@earendil-works/pi-ai": "^0.87.1",
-    "@pi-unipi/core": "3.0.0-alpha.22",
+    "@pi-unipi/core": "3.0.0-alpha.23",
     "turndown": "^7.2.4"
   },
   "devDependencies": {

package/packages/background-tasks/src/config.ts

@@ -21,8 +21,6 @@ export interface BackgroundTasksConfig {
   notifyOnCompletion: boolean;
   /** Default triggerOnCompletion (follow-up wake) for bg_run tasks. */
   triggerOnCompletion: boolean;
-  /** Default timeoutSeconds for shell tasks (0 = none). */
-  defaultTimeoutSeconds: number;
   /** Max finished tasks retained in memory. */
   maxFinishedTasks: number;
   /** Output cap in bytes before a task is killed+failed (default 20 MiB). */
@@ -33,7 +31,6 @@ export const DEFAULT_CONFIG: BackgroundTasksConfig = {
   enabled: true,
   notifyOnCompletion: true,
   triggerOnCompletion: true,
-  defaultTimeoutSeconds: 0,
   maxFinishedTasks: 30,
   maxOutputBytes: 20 * 1024 * 1024,
 };
@@ -114,11 +111,6 @@ export function validateBackgroundTasksConfig(config: unknown): string[] {
       problems.push(`${key} must be a boolean`);
     }
   }
-  if (c.defaultTimeoutSeconds !== undefined) {
-    if (typeof c.defaultTimeoutSeconds !== "number" || !Number.isFinite(c.defaultTimeoutSeconds) || c.defaultTimeoutSeconds < 0) {
-      problems.push("defaultTimeoutSeconds must be a non-negative number");
-    }
-  }
   if (c.maxFinishedTasks !== undefined) {
     if (typeof c.maxFinishedTasks !== "number" || !Number.isInteger(c.maxFinishedTasks) || c.maxFinishedTasks < 1) {
       problems.push("maxFinishedTasks must be a positive integer");
@@ -154,12 +146,11 @@ registerSettings({
     {
       title: "Tasks",
       fields: [
-        { key: "enabled", type: "boolean", label: "Enabled", description: "Master toggle — off registers nothing" },
-        { key: "notifyOnCompletion", type: "boolean", label: "Notify on completion" },
-        { key: "triggerOnCompletion", type: "boolean", label: "Follow-up wake", description: "Terminal state wakes the agent" },
-        { key: "defaultTimeoutSeconds", type: "number", label: "Timeout s", min: 0, zeroLabel: "∞ none" },
-        { key: "maxFinishedTasks", type: "number", label: "Max finished kept", min: 1 },
-        { key: "maxOutputBytes", type: "number", label: "Output cap bytes", min: 1024, description: "Kill+fail a task past this output size" },
+        { key: "enabled", type: "boolean", label: "Enabled", description: "Master toggle — off registers no tools, commands, hooks or UI." },
+        { key: "notifyOnCompletion", type: "boolean", label: "Notify on completion", description: "Default for tool-launched tasks; bg_run can override per call." },
+        { key: "triggerOnCompletion", type: "boolean", label: "Follow-up wake", description: "A finished task's terminal state wakes the agent for a follow-up turn." },
+        { key: "maxFinishedTasks", type: "number", label: "Max finished kept", min: 1, description: "Finished tasks kept in memory before the oldest is dropped." },
+        { key: "maxOutputBytes", type: "number", label: "Output cap", unit: "bytes", min: 1024, description: "Kill and fail a task that outputs more than this." },
       ],
     },
   ],
@@ -202,10 +193,6 @@ export function loadBackgroundTasksConfig(cwd: string): LoadedBackgroundTasksCon
       typeof merged.notifyOnCompletion === "boolean" ? merged.notifyOnCompletion : DEFAULT_CONFIG.notifyOnCompletion,
     triggerOnCompletion:
       typeof merged.triggerOnCompletion === "boolean" ? merged.triggerOnCompletion : DEFAULT_CONFIG.triggerOnCompletion,
-    defaultTimeoutSeconds:
-      typeof merged.defaultTimeoutSeconds === "number" && Number.isFinite(merged.defaultTimeoutSeconds) && merged.defaultTimeoutSeconds >= 0
-        ? merged.defaultTimeoutSeconds
-        : DEFAULT_CONFIG.defaultTimeoutSeconds,
     maxFinishedTasks:
       typeof merged.maxFinishedTasks === "number" && Number.isInteger(merged.maxFinishedTasks) && merged.maxFinishedTasks >= 1
         ? merged.maxFinishedTasks

package/packages/btw/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pi-unipi/btw",
-  "version": "3.0.0-alpha.22",
+  "version": "3.0.0-alpha.23",
   "description": "A pi extension for parallel side conversations with /unipi:btw — part of the Unipi suite",
   "type": "module",
   "main": "extensions/btw.ts",
@@ -37,7 +37,7 @@
     "@earendil-works/pi-tui": "^0.87.1"
   },
   "dependencies": {
-    "@pi-unipi/core": "3.0.0-alpha.22"
+    "@pi-unipi/core": "3.0.0-alpha.23"
   },
   "pi": {
     "extensions": [

package/packages/compactor/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pi-unipi/compactor",
-  "version": "3.0.0-alpha.22",
+  "version": "3.0.0-alpha.23",
   "description": "Compaction for Pi — lossless zero-LLM or model summaries, rebuilt from full history, with session recall",
   "type": "module",
   "main": "src/index.ts",
@@ -33,8 +33,8 @@
     "access": "public"
   },
   "dependencies": {
-    "@pi-unipi/core": "3.0.0-alpha.22",
-    "@pi-unipi/info-screen": "3.0.0-alpha.22",
+    "@pi-unipi/core": "3.0.0-alpha.23",
+    "@pi-unipi/info-screen": "3.0.0-alpha.23",
     "@earendil-works/pi-agent-core": "^0.87.1"
   },
   "peerDependencies": {

package/packages/compactor/src/compaction/hooks.ts

@@ -362,7 +362,7 @@ export function registerCompactionHooks(pi: ExtensionAPI, deps: CompactionHookDe
       } catch (err) {
         if (event.signal?.aborted) return { cancel: true };
         // Fall back to Pi's own path (same summarizer, Pi's auth handling).
-        dbg(config.debug, { method: "llm", fallback: true, error: String(err) });
+        dbg(process.env.UNIPI_COMPACTOR_DEBUG === "1", { method: "llm", fallback: true, error: String(err) });
         return;
       }
     }
@@ -394,14 +394,14 @@ export function registerCompactionHooks(pi: ExtensionAPI, deps: CompactionHookDe
           ctx?.ui?.notify?.(plan.reason === "no_live_messages" ? "compactor: nothing to compact" : "compactor: too few messages to compact", "warning");
         } catch {}
       }
-      dbg(config.debug, { cancelled: true, reason: plan.reason });
+      dbg(process.env.UNIPI_COMPACTOR_DEBUG === "1", { cancelled: true, reason: plan.reason });
       return { cancel: true };
     }
 
     lastStats = plan.stats;
     lastDetails = plan.details;
     if (!parsed.isCompactor && parsed.followUpPrompt && reason === "manual") pendingFollowUpPrompt = parsed.followUpPrompt;
-    dbg(config.debug, { method: "vcc", reason, stats: plan.stats, details: plan.details, summaryPreview: plan.summary.slice(0, 800) });
+    dbg(process.env.UNIPI_COMPACTOR_DEBUG === "1", { method: "vcc", reason, stats: plan.stats, details: plan.details, summaryPreview: plan.summary.slice(0, 800) });
     return {
       compaction: {
         summary: plan.summary,

package/packages/compactor/src/config/manager.ts

@@ -6,11 +6,24 @@ import { homedir } from "node:os";
 import { join } from "node:path";
 import { getSettingsScoped, registerSettings, setSettings } from "@pi-unipi/core";
 import type { CompactorConfig } from "../types.js";
-import { DEFAULT_COMPACTOR_CONFIG } from "./schema.js";
+import { DEFAULT_COMPACTOR_CONFIG, SUMMARY_SECTION_IDS } from "./schema.js";
 
 const METHOD_OPTIONS = [
-  { value: "vcc", label: "lossless (no model)" },
-  { value: "llm", label: "model summary" },
+  { value: "vcc", label: "lossless", description: "instant structured summary; history stays searchable" },
+  { value: "llm", label: "model summary", description: "pi's model writes the summary — costs one model call" },
+] as const;
+
+/** The `summarySections` multiselect: labels + explanations per section id. */
+const SUMMARY_SECTION_OPTIONS = [
+  { value: "activeWork", label: "Active work", description: "goal, ralph and kanboard state" },
+  { value: "requests", label: "Your requests", description: "the user requests this session answered" },
+  { value: "state", label: "Latest state", description: "the agent's most recent progress reports" },
+  { value: "decisions", label: "Decisions", description: "decisions and constraints reached so far" },
+  { value: "files", label: "Files", description: "files touched and what was done to them" },
+  { value: "commits", label: "Commits", description: "commit history of this session" },
+  { value: "errors", label: "Open errors", description: "unresolved errors worth remembering" },
+  { value: "lessons", label: "Lessons", description: "memory notes and diagnoses written down" },
+  { value: "transcript", label: "Recent transcript", description: "the last turns kept verbatim" },
 ] as const;
 
 // Registered with the unified settings hub. Canonical paths:
@@ -29,28 +42,28 @@ registerSettings({
           key: "method",
           type: "enum",
           label: "Method",
-          description: "Lossless: instant structured summary, full history stays searchable. Model summary: Pi's model-written summary (costs a model call).",
+          description: "How the context is summarized. Takes effect on the next compaction.",
           options: METHOD_OPTIONS,
         },
         {
           key: "piCompact",
           type: "enum",
           label: "Pi's /compact",
-          description: "What Pi's built-in /compact command does",
-          options: [{ value: "follow", label: "same as Method" }, ...METHOD_OPTIONS],
+          description: "What pi's built-in /compact command does: follow Method, or force one.",
+          options: [{ value: "follow", label: "same as Method", description: "reuse whatever Method is set to" }, ...METHOD_OPTIONS],
         },
         {
           key: "trigger",
           type: "enum",
           label: "When",
-          description: "Pi's limit: compact when the context nears the model's window (Pi's compaction settings). Percentage: compact at a set % of the window.",
+          description: "What trips a compaction while you work.",
           options: [
-            { value: "pi", label: "Pi's context limit" },
-            { value: "percent", label: "at a percentage" },
+            { value: "pi", label: "pi's context limit", description: "compact when pi itself nears the model's window" },
+            { value: "percent", label: "at a percentage", description: "compact when usage crosses the threshold below" },
           ],
         },
-        { key: "thresholdPercent", type: "number", label: "Percentage", min: 30, max: 95, description: "Used when When = at a percentage" },
-        { key: "notify", type: "boolean", label: "Notifications", description: "Show a notice when compaction runs or fails" },
+        { key: "thresholdPercent", type: "number", label: "Trigger threshold", unit: "%", min: 30, max: 95, description: "Context share that trips a compaction when When = at a percentage." },
+        { key: "notify", type: "boolean", label: "Notifications", description: "Notice when a compaction runs or fails." },
       ],
     },
     {
@@ -58,21 +71,17 @@ registerSettings({
       description: "Tuned defaults — rarely worth changing",
       advanced: true,
       fields: [
-        { key: "smartKeepTail", type: "boolean", label: "Smart keep tail", description: "Keep more recent turns when the kept tail would be tiny (≤5k tokens, up to 25k)" },
-        { key: "summaryBudgetTokens", type: "number", label: "Summary budget", min: 0, max: 20000, zeroLabel: "auto", description: "Lossless summary size in tokens (auto scales 1.5k–4k with session size)" },
-        { key: "sections.activeWork", type: "boolean", label: "Section: active work", description: "Goal, ralph and kanboard state from those modules" },
-        { key: "sections.requests", type: "boolean", label: "Section: your requests" },
-        { key: "sections.state", type: "boolean", label: "Section: latest state", description: "The agent's most recent progress reports" },
-        { key: "sections.decisions", type: "boolean", label: "Section: decisions & constraints" },
-        { key: "sections.files", type: "boolean", label: "Section: files" },
-        { key: "sections.commits", type: "boolean", label: "Section: commits" },
-        { key: "sections.errors", type: "boolean", label: "Section: open errors" },
-        { key: "sections.lessons", type: "boolean", label: "Section: lessons", description: "Lessons the agent wrote down: memory notes, # comments, diagnoses" },
-        { key: "sections.transcript", type: "boolean", label: "Section: recent transcript" },
-        { key: "cooldownMs", type: "number", label: "Percentage cooldown ms", min: 0, zeroLabel: "none", description: "Minimum delay between percentage-triggered compactions" },
-        { key: "repeatMinGrowthTokens", type: "number", label: "Percentage repeat growth", min: 0, zeroLabel: "off", description: "New tokens needed to compact again while still above the percentage" },
-        { key: "llmInstructions", type: "string", label: "Model summary instructions", emptyLabel: "none", description: "Extra instructions passed to model-written summaries" },
-        { key: "debug", type: "boolean", label: "Debug output", description: "Write compaction diagnostics to /tmp/compactor-debug.json" },
+        { key: "smartKeepTail", type: "boolean", label: "Smart keep tail", description: "Keep more recent turns when the kept tail would be tiny (≤5k tokens, up to 25k)." },
+        { key: "summaryBudgetTokens", type: "number", label: "Summary budget", unit: "tokens", min: 0, max: 20000, zeroLabel: "auto", description: "Lossless summary size in tokens (auto scales 1.5k–4k with session size)." },
+        {
+          key: "summarySections",
+          type: "multiselect",
+          label: "Summary sections",
+          description: "What the summary keeps — deselect to drop it from every new summary.",
+          options: SUMMARY_SECTION_OPTIONS,
+          emptyLabel: "none",
+        },
+        { key: "llmInstructions", type: "string", label: "Summary instructions", emptyLabel: "none", description: "Extra instructions passed to model-written summaries." },
       ],
     },
   ],
@@ -103,8 +112,6 @@ export function translateLegacyConfig(raw: Raw): Raw {
     if (!("trigger" in raw) && auto.enabled === true) out.trigger = "percent";
     const carry: Array<[keyof CompactorConfig, string]> = [
       ["thresholdPercent", "thresholdPercent"],
-      ["cooldownMs", "cooldownMs"],
-      ["repeatMinGrowthTokens", "repeatMinGrowthTokens"],
       ["notify", "notify"],
     ];
     for (const [to, from] of carry) {
@@ -137,9 +144,49 @@ export function loadConfig(cwd: string = process.cwd()): CompactorConfig {
   }
   if (config.method === "jev") config.method = "vcc";
   if (config.piCompact === "jev") config.piCompact = "vcc";
+  // The defaults-merged object always carries summarySections, so the stored
+  // layers decide: new array wins, else legacy sections booleans, else default.
+  const selection = storedSummarySelection(cwd);
+  if (selection !== undefined) config.summarySections = selection;
+  normalizeSummarySections(config);
   return config as unknown as CompactorConfig;
 }
 
+/**
+ * The stored section selection: first layer (project wins) with the new
+ * array, else the first with a legacy sections object (membership = every id
+ * not explicitly false). undefined = nothing stored — the default applies.
+ */
+function storedSummarySelection(cwd: string): string[] | undefined {
+  for (const scope of ["project", "global"] as const) {
+    let layer: Raw | undefined;
+    try {
+      layer = getSettingsScoped("compactor", scope, cwd);
+    } catch {
+      layer = undefined;
+    }
+    if (!layer) continue;
+    if (Array.isArray(layer.summarySections)) return layer.summarySections.map(String);
+    if (isRecord(layer.sections)) {
+      const legacy = layer.sections;
+      return SUMMARY_SECTION_IDS.filter((id) => legacy[id] !== false);
+    }
+  }
+  return undefined;
+}
+
+/**
+ * `summarySections` (string[]) present → the internal `sections` object
+ * derives from membership; absent → the default selection stands.
+ */
+function normalizeSummarySections(config: Raw): void {
+  if (!Array.isArray(config.summarySections)) return;
+  const on = new Set(config.summarySections.map(String));
+  const sections: Raw = {};
+  for (const id of SUMMARY_SECTION_IDS) sections[id] = on.has(id);
+  config.sections = sections;
+}
+
 /**
  * Write translated legacy keys into each scope's file once, so the settings
  * hub shows the effective values. Additive only: old keys stay, untouched.
@@ -148,10 +195,18 @@ export function migrateLegacyConfigFiles(cwd: string = process.cwd()): void {
   for (const scope of ["global", "project"] as const) {
     try {
       const raw = getSettingsScoped("compactor", scope, cwd);
-      if (!raw || (!("overrideDefaultCompaction" in raw) && !("autoCompaction" in raw))) continue;
+      if (!raw) continue;
+      // summarySections: derive from a stored legacy sections object once, so
+      // the hub shows the effective selection. Additive — booleans stay.
+      if (!Array.isArray(raw.summarySections) && isRecord(raw.sections)) {
+        const derived = SUMMARY_SECTION_IDS.filter((id) => (raw.sections as Raw)[id] !== false);
+        setSettings("compactor", { summarySections: [...derived] }, scope, cwd);
+        raw.summarySections = derived;
+      }
+      if (!("overrideDefaultCompaction" in raw) && !("autoCompaction" in raw)) continue;
       const translated = translateLegacyConfig(raw);
       const patch: Raw = {};
-      for (const key of ["method", "trigger", "thresholdPercent", "cooldownMs", "repeatMinGrowthTokens", "notify"]) {
+      for (const key of ["method", "trigger", "thresholdPercent", "notify", "summarySections"]) {
         if (!(key in raw) && key in translated) patch[key] = translated[key];
       }
       if (Object.keys(patch).length > 0) setSettings("compactor", patch, scope, cwd);

package/packages/compactor/src/config/schema.ts

@@ -4,6 +4,23 @@
 
 import type { AutoCompactionConfig, CompactorConfig } from "../types.js";
 
+/** Percentage-trigger tuning, fixed since the settings were pruned (UNI-50). */
+export const COOLDOWN_MS = 60_000;
+export const REPEAT_MIN_GROWTH_TOKENS = 4_000;
+
+/** Summary-section ids in canonical order (hub multiselect `summarySections`). */
+export const SUMMARY_SECTION_IDS = [
+  "activeWork",
+  "requests",
+  "state",
+  "decisions",
+  "files",
+  "commits",
+  "errors",
+  "lessons",
+  "transcript",
+] as const;
+
 export const DEFAULT_COMPACTOR_CONFIG: CompactorConfig = {
   method: "vcc",
   piCompact: "follow",
@@ -23,10 +40,8 @@ export const DEFAULT_COMPACTOR_CONFIG: CompactorConfig = {
     lessons: true,
     transcript: true,
   },
-  cooldownMs: 60_000,
-  repeatMinGrowthTokens: 4_000,
+  summarySections: [...SUMMARY_SECTION_IDS],
   llmInstructions: "",
-  debug: false,
 };
 
 /** The percentage-trigger view of a config (input to decideAutoCompaction). */
@@ -34,8 +49,8 @@ export function autoCompactionOf(config: CompactorConfig): AutoCompactionConfig
   return {
     enabled: config.trigger === "percent",
     thresholdPercent: config.thresholdPercent,
-    cooldownMs: config.cooldownMs,
-    repeatMinGrowthTokens: config.repeatMinGrowthTokens,
+    cooldownMs: COOLDOWN_MS,
+    repeatMinGrowthTokens: REPEAT_MIN_GROWTH_TOKENS,
     notify: config.notify,
   };
 }

package/packages/compactor/src/types.ts

@@ -88,14 +88,10 @@ export interface CompactorConfig {
   /** Lossless summary budget in tokens; 0 = auto (scales with session size). */
   summaryBudgetTokens: number;
   sections: SummarySections;
-  /** Percentage trigger: minimum delay between compactions. */
-  cooldownMs: number;
-  /** Percentage trigger: new tokens needed to compact again while still above threshold. */
-  repeatMinGrowthTokens: number;
+  /** Hub multiselect form of `sections`; present → sections derives from it. */
+  summarySections?: string[];
   /** Extra instructions for model-written summaries. */
   llmInstructions: string;
-  /** Write compaction diagnostics to /tmp/compactor-debug.json. */
-  debug: boolean;
 }
 
 /** Percentage auto-compaction trigger settings (decision input). */

package/packages/core/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pi-unipi/core",
-  "version": "3.0.0-alpha.22",
+  "version": "3.0.0-alpha.23",
   "description": "Shared utilities, event types, and constants for Unipi extension suite",
   "type": "module",
   "license": "MIT",