dankvoice: humanize answers via the @tiny role and narrate thinking

Answers spoken with /say were verbatim markdown-cleaned text while the
built-in vocalizer speaks prose rewritten by a small model. Port that
enhancer (bundle Qyr/T6o: same prompt, 1536 tokens, 6 s, 4000-char
window, null-fallback/empty-skip contract) and resolve its model from
modelRoles.tiny exactly as the harness does, so auto-speak sounds like
the enhanced TUI voice instead of a pager. agent_end now also speaks the
last assistant message's thinking blocks as their own unit before the
answer; narration-flagged thinking counts as answer text. Both knobs are
menu-toggleable, persisted, and env-overridable; the fixture gains a
chat route (echo/prose) and seven scenarios cover the round-trip, the
404 fallback, ordering, the kill switches and role resolution (22 green,
58 suite-wide; live smoke through real frank-zealot + frank-tts).

Refs #27

agent/config.yml

@@ -1,6 +1,10 @@
 providers:
   openrouterVariant: online
   fetch: auto
+  maxInFlightRequests:
+    llamacpp (dankserver): 3
+    llamacpp (dankclaw): 2
+    llamacpp (dankclaw-3090): 3
 modelRoles:
   slow: llamacpp (dankclaw)/frank-dragoon:medium
   vision: llamacpp (dankclaw-3090)/frank-zealot:xhigh
@@ -145,17 +149,24 @@ display:
   subagentLivePreview: true
 task:
   showResolvedModelBadge: true
-  maxConcurrency: 3
+  maxConcurrency: 2
   agentModelOverrides:
     designer: llamacpp (dankclaw-3090)/frank-zealot
     librarian: llamacpp (dankclaw-3090)/frank-zealot
     reviewer: llamacpp (dankserver)/frank-archon
-    scout: llamacpp (dankserver)/frank-archon
+    scout: llamacpp (dankclaw-3090)/frank-zealot
     security-reviewer: llamacpp (dankclaw)/frank-dragoon
     sonic: llamacpp (dankclaw-3090)/frank-zealot
     task: llamacpp (dankserver)/frank-archon
   eager: default
   enableLsp: true
+  maxRuntimeMs: 1800000
+  softRequestBudget: 120
+  maxRecursionDepth: 1
+  disabledAgents: []
+  isolation:
+    enabled: true
+  completionProbe: false
 snapcompact:
   systemPrompt: none
   toolResults: false
@@ -203,7 +214,7 @@ vault:
 tts:
   localVoice: bm_george
 speech:
-  enabled: true
+  enabled: false
   mode: all
   enhanced: true
   voice: am_fenrir
@@ -232,6 +243,7 @@ followUpMode: one-at-a-time
 tools:
   approvalMode: yolo
   format: auto
+  maxTimeout: 600
 marketplace:
   autoUpdate: notify
 bash:

agent/extensions/dankvoice.ts

@@ -7,28 +7,47 @@
  * agent/models.yml) could only ever produce files. This extension adds the
  * missing path — the TUI speaks with the cloned voices:
  *
- *   /say <text>          speak once through the speakers
- *   /say on | off        auto-speak the assistant's final answer each turn
- *   /say voice <name>    choose a voice (empty = whatever the server defaults to)
- *   /say voices          list the voices frank-tts reports
- *   /say stop            cut off the current utterance
- *   /say status          voice, endpoint and player in use
+ *   /say                   open the options menu
+ *   /say <text>            speak once through the speakers
+ *   /say on | off          arm/disarm auto-speak of the assistant's final answer
+ *   /say menu              the options menu (voice, output device, player, cap)
+ *   /say voice <name>      choose a voice (empty = whatever the server defaults to)
+ *   /say voices            list the voices frank-tts reports
+ *   /say stop              cut off the current utterance
+ *   /say status            voice, endpoint, player and device in use
+ *   /say help              the command list
  *
  * The endpoint comes from agent/models.yml — the provider block carrying
  * `api: openai-speech`, i.e. the same row the speech role resolves — so this
- * file carries no host address and no secret. Environment overrides:
+ * file carries no host address and no secret. Menu choices persist to
+ * ~/.omp/cache/dankvoice.json (runtime state, never tracked). Environment
+ * overrides win over the file, which wins over defaults:
  *
  *   DANKVOICE_URL      speech base URL (…/v1) instead of models.yml
  *   DANKVOICE_MODEL    served model id (default: the id in that block)
- *   DANKVOICE_VOICE    default voice name
- *   DANKVOICE_AUTO=1   start with auto-speak armed
- *   DANKVOICE_MAX      spoken-character cap per utterance (default 700)
+ *   DANKVOICE_CONFIG   settings file (default ~/.omp/cache/dankvoice.json)
  *   DANKVOICE_PLAYER   player binary, skipping detection
- *   DANKVOICE_MODELS   path to models.yml (default ~/.omp/agent/models.yml)
+ *   DANKVOICE_VOICE    voice name ("" = server default)
+ *   DANKVOICE_MAX      spoken-character budget per answer (default 700)
+ *   DANKVOICE_AUTO=1   start with auto-speak armed
+ *   DANKVOICE_REWRITE_URL   @tiny chat base URL (…/v1) for spoken-prose rewrite
+ *   DANKVOICE_REWRITE_MODEL chat model id (default: the @tiny role's model)
+ *   DANKVOICE_REWRITE=0     disable the rewrite step (speak mechanical text)
+ *   DANKVOICE_THINKING=0|1  narration of thinking blocks (default: on)
+ *   DANKVOICE_OMPCONFIG     omp config.yml to resolve the @tiny role from
  *
- * Utterances are synthesized to a temp WAV in $TMPDIR; the file is removed
- * when its player process exits. A new utterance interrupts the previous one
- * instead of queueing behind it.
+ * Long answers are cut into speakable segments by a state machine ported
+ * from omp's built-in vocalizer: strong boundaries on [.!?…] runs followed
+ * by whitespace (so "!!!!!" and "..." stay one boundary), guards for
+ * abbreviations and decimals, code fences and tables swallowed, markdown
+ * cleaned. Each segment is synthesized to a temp WAV in $TMPDIR and played
+ * to completion before the next request goes out — requests are strictly
+ * sequential, so audio never stacks. The chain runs detached from the event
+ * handler, so an utterance may outlive the harness handler timeout. A new
+ * utterance interrupts the previous one. Output routing: a saved sink
+ * becomes `pw-play --target <sink>` or `paplay --device=<sink>` (ffplay has
+ * no device concept and ignores it); with no saved sink the player's own
+ * default applies.
  */
 
 import { spawn, spawnSync, type ChildProcess } from "node:child_process";
@@ -62,6 +81,7 @@ export interface Player {
 }
 
 export type SayAction =
+  | { kind: "menu" }
   | { kind: "help" }
   | { kind: "speak"; text: string }
   | { kind: "on" }
@@ -75,7 +95,8 @@ export type SayAction =
 export function parseSayArgs(raw: string): SayAction {
   const args = raw.trim();
   const lower = args.toLowerCase();
-  if (lower === "" || lower === "help") return { kind: "help" };
+  if (lower === "" || lower === "menu") return { kind: "menu" };
+  if (lower === "help") return { kind: "help" };
   if (lower === "on") return { kind: "on" };
   if (lower === "off") return { kind: "off" };
   if (lower === "stop") return { kind: "stop" };
@@ -162,38 +183,435 @@ export function parseVoices(payload: unknown): string[] {
   return names.filter((name) => name.length > 0);
 }
 
-/** Player preference: PipeWire first (this fleet runs WirePlumber), then Pulse, then ffmpeg. */
-export function detectPlayer(lookup: (command: string) => string | null): Player | null {
-  const candidates: Player[] = [
-    { command: "pw-play", args: [] },
-    { command: "paplay", args: [] },
-    { command: "ffplay", args: ["-nodisp", "-autoexit", "-loglevel", "quiet"] },
-  ];
-  for (const candidate of candidates) {
-    if (lookup(candidate.command)) return candidate;
-  }
-  return null;
+/** Known players, PipeWire first (the workstations run WirePlumber); first existing wins. */
+export function availablePlayers(lookup: (command: string) => string | null): string[] {
+  return ["pw-play", "paplay", "ffplay"].filter((command) => lookup(command) !== null);
 }
 
-/** What a speech engine should say: prose only, no fences, no markup, bounded. */
-export function stripForSpeech(text: string, maxChars: number = DEFAULT_MAX_CHARS): string {
-  let out = text;
-  out = out.replace(/```[\s\S]*?```/g, " ");
-  out = out.replace(/~~~[\s\S]*?~~~/g, " ");
-  out = out.replace(/`([^`]*)`/g, "$1");
-  out = out.replace(/!\[[^\]]*\]\([^)]*\)/g, " ");
-  out = out.replace(/\[([^\]]*)\]\([^)]*\)/g, "$1");
-  out = out.replace(/^\s{0,3}#{1,6}\s+/gm, "");
-  out = out.replace(/^\s*[-*+]\s+/gm, "");
-  out = out.replace(/^\s*\|[-: |]+\|\s*$/gm, " ");
-  out = out.replace(/\|/g, " ");
-  out = out.replace(/[*_]{1,3}([^*_]+)[*_]{1,3}/g, "$1");
-  out = out.replace(/\s+/g, " ").trim();
-  if (out.length <= maxChars) return out;
-  const window = out.slice(0, maxChars);
-  const cut = Math.max(window.lastIndexOf(". "), window.lastIndexOf("! "), window.lastIndexOf("? "));
-  const clipped = cut > maxChars / 2 ? window.slice(0, cut + 1) : window.replace(/[,;:]\s*\S*$/, "");
-  return `${clipped.trim()} …`;
+/** One player invocation: the saved sink becomes each player's routing dialect. */
+export function playerFor(command: string, sink: string): Player {
+  const base = command.split("/").pop() ?? command;
+  if (base === "ffplay") return { command, args: ["-nodisp", "-autoexit", "-loglevel", "quiet"] };
+  if (!sink) return { command, args: [] };
+  if (base === "paplay") return { command, args: [`--device=${sink}`] };
+  return { command, args: ["--target", sink] };
+}
+
+// ------------------------------------------------------------ segmenting
+//
+// Ported from omp's built-in speak-aloud vocalizer (the `fHe` segmenter and
+// its helpers in dist/cli.js). The same boundary rules as the stock path,
+// kept deliberately: push() takes text and returns the segments that are
+// complete, flush() ends the message.
+
+const SENTENCE_END = /[.!?\u2026]+[)\]"'\u00BB\u201D\u2019]*\s/g;
+const WEAK_BREAK = /[,;:\u2014\u2013]\s/g;
+const ABBREVIATION = /(?:^|\s)(?:e\.g|i\.e|etc|vs|Mr|Mrs|Ms|Dr|St|No)\.$/i;
+const LINE_HEAD = /^(?:#{1,6}|[-*+]|-{2,}|\*{2,}|_{2,}|\d{1,3}|\d{1,3}[.)]|>+|`{1,2}|~{1,2})$/;
+const HORIZONTAL_RULE = /^(?:-{3,}|\*{3,}|_{3,})\s*$/;
+
+const IMAGE_LINK = /!\[([^\]]*)\]\(([^()]*)\)/g;
+const LINK = /\[([^\]]+)\]\(([^()]*)\)/g;
+const BRACKETED_URL = /<(https?:\/\/[^\s>]+)>/g;
+const BARE_URL = /\bhttps?:\/\/[^\s<>()"'\]]+|\bwww\.[\w-]+(?:\.[\w-]+)+[^\s<>()"'\]]*/g;
+const INLINE_CODE = /`{1,2}([^`]+)`{1,2}/g;
+const BOLD_MARKUP = /\*\*|__|~~/g;
+const STAR = /\*(?=\S)|(?<=\S)\*/g;
+const UNDERSCORE = /(^|\s)_+|_+(?=\s|$)/g;
+const HTML_TAG = /<\/?[a-zA-Z][^<>]*>/g;
+const EMPHASIS_RUN = /(^|\s)[-*_]{3,}(?=\s|$)/g;
+const DEEP_PATH = /(^|[\s("'`])((?:~|\.{1,2})?\/?[\w.@+-]+(?:\/[\w.@+-]+){2,}\/?)/g;
+const HAS_WORD = /[\p{L}\p{N}]/u;
+
+/** Speak a URL by host and a file path by its last segment. */
+export function spokenDomain(url: string): string {
+  return url.replace(/^[a-z][\w+.-]*:\/\//i, "").replace(/^www\./i, "").replace(/[/?#].*$/, "");
+}
+
+/** Final per-segment cleanup; empty when nothing speakable survived. */
+export function cleanForSpeech(text: string): string {
+  const out = text
+    .replace(IMAGE_LINK, "$1")
+    .replace(LINK, "$1")
+    .replace(BRACKETED_URL, (_match, url: string) => spokenDomain(url))
+    .replace(BARE_URL, (url: string) => spokenDomain(url))
+    .replace(INLINE_CODE, "$1")
+    .replace(BOLD_MARKUP, "")
+    .replace(STAR, "")
+    .replace(UNDERSCORE, "$1")
+    .replace(HTML_TAG, " ")
+    .replace(EMPHASIS_RUN, "$1")
+    .replace(DEEP_PATH, (_match, pre: string, path: string) => {
+      const parts = path.split("/").filter((part) => part.length > 0);
+      return pre + (parts[parts.length - 1] ?? path);
+    })
+    .replace(/\s+/g, " ")
+    .trim();
+  return HAS_WORD.test(out) ? out : "";
+}
+
+function sentenceBoundary(text: string, minChars: number): number {
+  SENTENCE_END.lastIndex = 0;
+  for (let match = SENTENCE_END.exec(text); match; match = SENTENCE_END.exec(text)) {
+    const end = match.index + match[0].length;
+    if (end < minChars) continue;
+    if (ABBREVIATION.test(text.slice(0, end).trimEnd())) continue;
+    if ((text.slice(0, end).match(/`/g)?.length ?? 0) % 2 !== 0) continue;
+    return end;
+  }
+  return -1;
+}
+
+function firstWeakBreak(text: string, minChars: number): number {
+  WEAK_BREAK.lastIndex = 0;
+  for (let match = WEAK_BREAK.exec(text); match; match = WEAK_BREAK.exec(text)) {
+    const end = match.index + match[0].length;
+    if (end >= minChars) return end;
+  }
+  return -1;
+}
+
+function lastWeakBreakWithin(text: string, minChars: number, maxChars: number): number {
+  WEAK_BREAK.lastIndex = 0;
+  let found = -1;
+  for (let match = WEAK_BREAK.exec(text); match; match = WEAK_BREAK.exec(text)) {
+    const end = match.index + match[0].length;
+    if (end > maxChars) break;
+    if (end >= minChars) found = end;
+  }
+  return found;
+}
+
+function wordWrapCut(text: string, maxChars: number): number {
+  const space = text.lastIndexOf(" ", maxChars);
+  return space > 0 ? space + 1 : Math.min(maxChars, text.length);
+}
+
+type LineHead =
+  | { kind: "swallow" }
+  | { kind: "fence"; fence: string }
+  | { kind: "marker"; spoken: string }
+  | { kind: "prose"; text: string }
+  | { kind: "undecided" };
+
+/** Classify a line start: tables and fences swallow, list markers get spoken. */
+export function classifyLineHead(head: string): LineHead {
+  if (head === "|") return { kind: "swallow" };
+  if (/^(?:`{3}|~{3})/.test(head)) return { kind: "fence", fence: head.slice(0, 3) };
+  if (/^#{1,6}[ \t]/.test(head)) return { kind: "marker", spoken: "" };
+  if (/^[-*+][ \t]/.test(head)) return { kind: "marker", spoken: "" };
+  const ordered = /^(\d{1,3})[.)][ \t]/.exec(head);
+  if (ordered) return { kind: "marker", spoken: `${ordered[1]}, ` };
+  if (/^>+/.test(head) && !/^>+$/.test(head)) return { kind: "prose", text: head.replace(/^>+[ \t]?/, "") };
+  if (LINE_HEAD.test(head)) return { kind: "undecided" };
+  return { kind: "prose", text: head };
+}
+
+type SegmenterState = "linestart" | "prose" | "swallow" | "code";
+
+/** The built-in vocalizer's segmenter: streamed text in, speakable segments out. */
+export class SpeechSegmenter {
+  #state: SegmenterState = "linestart";
+  #head = "";
+  #fence = "";
+  #fenceLine = "";
+  #resume: SegmenterState = "linestart";
+  #buffer = "";
+  #started = false;
+
+  push(chunk: string): string[] {
+    const out: string[] = [];
+    for (const char of chunk) this.#feed(char, out);
+    this.#split(out);
+    return out;
+  }
+
+  flush(): string[] {
+    const out: string[] = [];
+    if (this.#state === "linestart" && this.#head.length > 0 && !HORIZONTAL_RULE.test(this.#head)) this.#buffer += this.#head;
+    this.#head = "";
+    this.#state = "linestart";
+    this.#emit(out);
+    return out;
+  }
+
+  /** Streaming idle flush: only cut a tail that already looks finished. */
+  flushIdle(): string[] {
+    const out: string[] = [];
+    const tail = this.#buffer.trimEnd();
+    if (!/[.!?\u2026][)\]"'\u00BB\u201D\u2019]*$/.test(tail) && tail.length < 24) return out;
+    this.#emit(out);
+    return out;
+  }
+
+  #feed(char: string, out: string[]): void {
+    switch (this.#state) {
+      case "linestart":
+        this.#feedLineStart(char, out);
+        return;
+      case "prose":
+        if (char === "\n") this.#endLine(out);
+        else this.#buffer += char;
+        return;
+      case "swallow":
+        if (char === "\n") this.#state = this.#resume;
+        return;
+      case "code":
+        this.#feedCode(char);
+        return;
+    }
+  }
+
+  #feedLineStart(char: string, out: string[]): void {
+    if (char === "\n") {
+      const head = this.#head;
+      this.#head = "";
+      if (head.length > 0 && !HORIZONTAL_RULE.test(head)) this.#buffer += head;
+      this.#endLine(out);
+      return;
+    }
+    this.#head += char;
+    const classified = classifyLineHead(this.#head);
+    if (classified.kind === "undecided") {
+      if (this.#head.length > 8) {
+        this.#buffer += this.#head;
+        this.#head = "";
+        this.#state = "prose";
+      }
+      return;
+    }
+    this.#head = "";
+    switch (classified.kind) {
+      case "prose":
+        this.#buffer += classified.text;
+        this.#state = "prose";
+        return;
+      case "marker":
+        this.#buffer += classified.spoken;
+        this.#state = "prose";
+        return;
+      case "swallow":
+        this.#state = "swallow";
+        this.#resume = "linestart";
+        return;
+      case "fence":
+        this.#fence = classified.fence;
+        this.#fenceLine = "";
+        this.#state = "swallow";
+        this.#resume = "code";
+        return;
+    }
+  }
+
+  #feedCode(char: string): void {
+    if (char === "\n") {
+      this.#fenceLine = "";
+      return;
+    }
+    if (this.#fenceLine.length < 3) {
+      this.#fenceLine += char;
+      if (this.#fenceLine === this.#fence) {
+        this.#state = "swallow";
+        this.#resume = "linestart";
+      }
+    }
+  }
+
+  #endLine(out: string[]): void {
+    this.#state = "linestart";
+    this.#emit(out);
+  }
+
+  #emit(out: string[]): void {
+    this.#split(out);
+    const segment = this.#buffer;
+    this.#buffer = "";
+    this.#pushSegment(segment, out);
+  }
+
+  #split(out: string[]): void {
+    for (;;) {
+      const buffer = this.#buffer;
+      const boundary = sentenceBoundary(buffer, this.#started ? 24 : 12);
+      if (boundary !== -1 && boundary <= 280) {
+        this.#cut(boundary, out);
+        continue;
+      }
+      if (!this.#started && buffer.length >= 40) {
+        const weak = firstWeakBreak(buffer, 12);
+        if (weak !== -1 && weak <= 140) {
+          this.#cut(weak, out);
+          continue;
+        }
+        if (buffer.length >= 140) {
+          this.#cut(wordWrapCut(buffer, 140), out);
+          continue;
+        }
+      }
+      if (this.#started && buffer.length >= 160) {
+        const weak = lastWeakBreakWithin(buffer, 24, 160);
+        if (weak !== -1) {
+          this.#cut(weak, out);
+          continue;
+        }
+      }
+      if (buffer.length > 280) {
+        const weak = lastWeakBreakWithin(buffer, 24, 280);
+        this.#cut(weak !== -1 ? weak : wordWrapCut(buffer, 280), out);
+        continue;
+      }
+      return;
+    }
+  }
+
+  #cut(end: number, out: string[]): void {
+    const head = this.#buffer.slice(0, end);
+    this.#buffer = this.#buffer.slice(end);
+    this.#pushSegment(head, out);
+  }
+
+  #pushSegment(raw: string, out: string[]): void {
+    const cleaned = cleanForSpeech(raw);
+    if (!cleaned) return;
+    out.push(cleaned);
+    this.#started = true;
+  }
+}
+
+/** A whole message to speakable segments: the segmenter run start to end. */
+export function segmentsForSpeech(text: string): string[] {
+  const segmenter = new SpeechSegmenter();
+  const segments = segmenter.push(text);
+  for (const tail of segmenter.flush()) segments.push(tail);
+  return segments;
+}
+
+/** The leading whole segments that fit the user's spoken-character budget. */
+export function budgetSegments(segments: string[], maxChars: number): string[] {
+  const chosen: string[] = [];
+  let used = 0;
+  for (const segment of segments) {
+    if (used + segment.length > maxChars) break;
+    used += segment.length;
+    chosen.push(segment);
+  }
+  return chosen;
+}
+
+// ------------------------------------------------------------ humanization
+//
+// Port of the built-in vocalizer's speech enhancer (the `Qyr` class and its
+// system prompt `T6o` in dist/cli.js): rewrite each unit into natural spoken
+// prose with the @tiny model before synthesis. Same prompt, same budgets
+// (1536 max tokens, 6 s, 4000-char input window), and the same contract:
+// null means the enhancer failed — fall back to the mechanical text — while
+// an empty string means the enhancer found nothing worth saying.
+
+const REWRITE_MAX_TOKENS = 1536;
+const REWRITE_TIMEOUT_MS = 6_000;
+const REWRITE_MAX_INPUT = 4_000;
+
+const REWRITE_SYSTEM_PROMPT = `Rewrite coding-assistant replies as words a person would say aloud. Audience: developer listening while working; written reply remains onscreen. Narrate; NEVER dictate syntax.
+
+Output ONLY spoken words—no Markdown, quotes, preamble, or stage directions.
+
+- Natural colleague-over-the-shoulder summary. Preserve original meaning, order, tone; add no opinions, greetings, or content absent from text.
+- NEVER read URLs, Markdown/table syntax, or separators. Link: label or site name ("the Bun issue on GitHub"). File path: file name only (say "vocalizer.ts," not "vocalizer dot t s").
+- Code blocks: NEVER read; replace each with one short clause describing it or its function ("a small helper that retries the request"). Skip it if surrounding prose already explains it.
+- Short inline identifiers, flags, commands: speak as-is ("run bun check"); paraphrase if awkward.
+- Speak numbers, versions, symbols naturally: "v1.2" → "version one point two"; "→" → "to"; "&" → "and"; "~5s" → "about five seconds".
+- Lists: flowing sentences ("first, then, and finally"); NEVER recite bullets or numbers.
+- Concise: same length or shorter; compress boilerplate, NEVER pad.
+- Partial mid-thought fragments: render only what exists; NEVER invent an ending.
+- Pure code, tables, or markup: empty reply.
+`;
+
+export interface RewriteTarget {
+  /** OpenAI-compatible chat completions route of the @tiny model. */
+  chatUrl: string;
+  /** Served model id to name in the request body. */
+  model: string;
+}
+
+/**
+ * Resolve the @tiny role the way the harness does: `modelRoles.tiny` in
+ * config.yml reads `provider (host)/model:effort`; the provider block in
+ * models.yml supplies the base URL. Unknown role or provider → undefined.
+ */
+export function resolveTinyRewriter(modelsYml: string, configYml: string): RewriteTarget | undefined {
+  const role = configYml.match(/^[ \t]*tiny:[ \t]*(.+?)[ \t]*$/m)?.[1];
+  if (!role) return undefined;
+  const ref = role.includes(":") ? role.slice(0, role.lastIndexOf(":")) : role;
+  const slash = ref.lastIndexOf("/");
+  if (slash <= 0) return undefined;
+  const wantProvider = ref.slice(0, slash).trim();
+  const wantModel = ref.slice(slash + 1).trim();
+  if (!wantProvider || !wantModel) return undefined;
+  let inProvider = false;
+  let base: string | undefined;
+  for (const line of modelsYml.split(/\r?\n/)) {
+    const key = line.match(/^ {2}"?([^"\s][^":]*?)"?[ \t]*:[ \t]*$/);
+    if (key) {
+      inProvider = key[1].trim() === wantProvider;
+      base = undefined;
+      continue;
+    }
+    if (!inProvider) continue;
+    const url = line.trim().match(/^baseUrl:\s*(\S+)/);
+    if (url) {
+      base = url[1].replace(/\/+$/, "");
+      continue;
+    }
+    if (line.trim().match(/^-\s*id:\s*(\S+)/)?.[1] === wantModel && base) {
+      return { chatUrl: `${base}/chat/completions`, model: wantModel };
+    }
+  }
+  return undefined;
+}
+
+/** Env override first, then the @tiny role; both miss → no humanization. */
+export function resolveRewriter(env: NodeJS.ProcessEnv, modelsYml: string, configYml: string): RewriteTarget | undefined {
+  const override = env.DANKVOICE_REWRITE_URL?.trim().replace(/\/+$/, "");
+  if (override) return { chatUrl: `${override}/chat/completions`, model: env.DANKVOICE_REWRITE_MODEL?.trim() || "tiny" };
+  return resolveTinyRewriter(modelsYml, configYml);
+}
+
+/** The enhancer's rewrite step; never throws — null is its "couldn't". */
+export async function rewriteForSpeech(target: RewriteTarget, text: string): Promise<string | null> {
+  try {
+    const res = await fetch(target.chatUrl, {
+      method: "POST",
+      headers: { "content-type": "application/json" },
+      body: JSON.stringify({
+        model: target.model,
+        messages: [
+          { role: "system", content: REWRITE_SYSTEM_PROMPT },
+          { role: "user", content: text.slice(0, REWRITE_MAX_INPUT) },
+        ],
+        max_tokens: REWRITE_MAX_TOKENS,
+      }),
+      signal: AbortSignal.timeout(REWRITE_TIMEOUT_MS),
+    });
+    if (!res.ok) return null;
+    const payload: unknown = await res.json();
+    if (!payload || typeof payload !== "object" || !("choices" in payload) || !Array.isArray(payload.choices) || payload.choices.length === 0) return null;
+    const first: unknown = payload.choices[0];
+    if (!first || typeof first !== "object" || !("message" in first)) return null;
+    const message: unknown = first.message;
+    if (!message || typeof message !== "object" || !("content" in message)) return null;
+    const content: unknown = message.content;
+    if (typeof content === "string") return content.trim();
+    if (!Array.isArray(content)) return null;
+    const parts: string[] = [];
+    for (const part of content) {
+      if (part && typeof part === "object" && "type" in part && part.type === "text" && "text" in part && typeof part.text === "string") {
+        parts.push(part.text);
+      }
+    }
+    return parts.join(" ").trim();
+  } catch {
+    return null;
+  }
 }
 
 /** The text the assistant last actually said; tool-only messages do not count. */
@@ -207,8 +625,12 @@ export function lastAssistantText(messages: readonly unknown[]): string {
     if (!Array.isArray(content)) continue;
     const parts: string[] = [];
     for (const part of content) {
-      if (part && typeof part === "object" && "type" in part && part.type === "text" && "text" in part && typeof part.text === "string") {
+      if (!part || typeof part !== "object" || !("type" in part)) continue;
+      if (part.type === "text" && "text" in part && typeof part.text === "string") {
         parts.push(part.text);
+      } else if (part.type === "thinking" && "narration" in part && part.narration === true && "thinking" in part && typeof part.thinking === "string") {
+        // provider-returned thinking addressed to the user: it is answer text
+        parts.push(part.thinking);
       }
     }
     const joined = parts.join("\n").trim();
@@ -217,14 +639,86 @@ export function lastAssistantText(messages: readonly unknown[]): string {
   return "";
 }
 
+/** The last assistant message's true thinking blocks; narration-flagged ones are answer text elsewhere. */
+export function lastAssistantThinking(messages: readonly unknown[]): string {
+  for (let i = messages.length - 1; i >= 0; i--) {
+    const message = messages[i];
+    if (!message || typeof message !== "object" || !("role" in message) || message.role !== "assistant") continue;
+    if (!("content" in message) || !Array.isArray(message.content)) return "";
+    const parts: string[] = [];
+    for (const part of message.content) {
+      if (part && typeof part === "object" && "type" in part && part.type === "thinking" && !(("narration" in part) && part.narration === true) && "thinking" in part && typeof part.thinking === "string") {
+        parts.push(part.thinking);
+      }
+    }
+    return parts.join("\n").trim();
+  }
+  return "";
+}
+
+/** Menu choices, persisted between sessions; every field optional. */
+export interface Settings {
+  voice?: string;
+  player?: string;
+  sink?: string;
+  maxChars?: number;
+  auto?: boolean;
+  thinking?: boolean;
+}
+
+/** Where menu choices live: DANKVOICE_CONFIG, else ~/.omp/cache/dankvoice.json. */
+export function settingsPath(env: NodeJS.ProcessEnv): string {
+  return env.DANKVOICE_CONFIG?.trim() || join(env.HOME ?? "", ".omp", "cache", "dankvoice.json");
+}
+
+export function readSettings(path: string): Settings {
+  try {
+    const parsed: unknown = JSON.parse(readFileSync(path, "utf8"));
+    if (!parsed || typeof parsed !== "object" || Array.isArray(parsed)) return {};
+    const raw = parsed as Record<string, unknown>;
+    const settings: Settings = {};
+    if (typeof raw.voice === "string") settings.voice = raw.voice;
+    if (typeof raw.player === "string") settings.player = raw.player;
+    if (typeof raw.sink === "string") settings.sink = raw.sink;
+    if (typeof raw.maxChars === "number" && Number.isFinite(raw.maxChars) && raw.maxChars > 0) settings.maxChars = Math.floor(raw.maxChars);
+    if (typeof raw.auto === "boolean") settings.auto = raw.auto;
+    if (typeof raw.thinking === "boolean") settings.thinking = raw.thinking;
+    return settings;
+  } catch {
+    return {};
+  }
+}
+
+/** A sink row from `wpctl list audio sinks`: `116<TAB>alsa_output…<TAB>audio/sink<TAB>*`. */
+export interface SinkRow {
+  id: string;
+  name: string;
+}
+
+export function parseSinkList(output: string): SinkRow[] {
+  const rows: SinkRow[] = [];
+  for (const line of output.split("\n")) {
+    const cols = line.split("\t").map((col) => col.trim());
+    if (cols.length < 3 || !cols[2]?.includes("audio/sink")) continue;
+    const id = cols[0] ?? "";
+    const name = cols[1] ?? "";
+    if (!id || !name) continue;
+    rows.push({ id, name });
+  }
+  return rows;
+}
+
 // ------------------------------------------------------------- live wiring
 
 interface Config {
   target?: SpeechTarget;
+  rewriter?: RewriteTarget;
   player?: Player;
+  sink: string;
   maxChars: number;
   auto: boolean;
   voice: string;
+  thinking: boolean;
 }
 
 function lookupOnPath(command: string): string | null {
@@ -241,10 +735,22 @@ function readModelsYml(): string {
   }
 }
 
+function readOmpConfigYml(): string {
+  const path = process.env.DANKVOICE_OMPCONFIG ?? join(process.env.HOME ?? "", ".omp", "agent", "config.yml");
+  try {
+    return readFileSync(path, "utf8");
+  } catch {
+    return "";
+  }
+}
+
 function loadConfig(): Config {
-  const overrideUrl = process.env.DANKVOICE_URL?.trim().replace(/\/+$/, "");
-  const modelOverride = process.env.DANKVOICE_MODEL?.trim();
-  const fromYml = overrideUrl ? undefined : resolveSpeechEndpoint(readModelsYml());
+  const env = process.env;
+  const settings = readSettings(settingsPath(env));
+  const modelsYml = readModelsYml();
+  const overrideUrl = env.DANKVOICE_URL?.trim().replace(/\/+$/, "");
+  const modelOverride = env.DANKVOICE_MODEL?.trim();
+  const fromYml = overrideUrl ? undefined : resolveSpeechEndpoint(modelsYml);
   const target: SpeechTarget | undefined = overrideUrl
     ? {
         speechUrl: `${overrideUrl}/audio/speech`,
@@ -255,17 +761,29 @@ function loadConfig(): Config {
     : fromYml
       ? { ...fromYml, model: modelOverride || fromYml.model }
       : undefined;
-  const playerOverride = process.env.DANKVOICE_PLAYER?.trim();
-  const maxChars = Number(process.env.DANKVOICE_MAX);
+  const playerCommand = env.DANKVOICE_PLAYER?.trim() || settings.player || availablePlayers(lookupOnPath)[0];
+  const sink = settings.sink ?? "";
+  const maxChars = Number(env.DANKVOICE_MAX);
+  const rewriter = env.DANKVOICE_REWRITE === "0" ? undefined : resolveRewriter(env, modelsYml, readOmpConfigYml());
+  const thinking = env.DANKVOICE_THINKING === "1" ? true : env.DANKVOICE_THINKING === "0" ? false : (settings.thinking ?? true);
   return {
     target,
-    player: playerOverride ? { command: playerOverride, args: [] } : (detectPlayer(lookupOnPath) ?? undefined),
-    maxChars: Number.isFinite(maxChars) && maxChars > 0 ? Math.floor(maxChars) : DEFAULT_MAX_CHARS,
-    auto: process.env.DANKVOICE_AUTO === "1",
-    voice: process.env.DANKVOICE_VOICE?.trim() ?? "",
+    rewriter,
+    player: playerCommand ? playerFor(playerCommand, sink) : undefined,
+    sink,
+    maxChars: Number.isFinite(maxChars) && maxChars > 0 ? Math.floor(maxChars) : (settings.maxChars ?? DEFAULT_MAX_CHARS),
+    auto: env.DANKVOICE_AUTO === "1" || (settings.auto ?? false),
+    voice: env.DANKVOICE_VOICE?.trim() ?? settings.voice ?? "",
+    thinking,
   };
 }
 
+/** PipeWire sinks via wpctl; empty when wireplumber is unreachable. */
+function listSinks(): SinkRow[] {
+  const probe = spawnSync("wpctl", ["list", "audio", "sinks"], { encoding: "utf8" });
+  return probe.status === 0 ? parseSinkList(probe.stdout) : [];
+}
+
 /** Synthesize one utterance into a WAV inside `dir`; throws with the endpoint's own message. */
 async function synthesize(target: SpeechTarget, body: Record<string, string>, dir: string, timeoutMs: number): Promise<string> {
   const res = await fetch(target.speechUrl, {
@@ -337,23 +855,65 @@ function startPlayback(player: Player, file: string, dir: string): ChildProcess
 
 export default function dankvoice(pi: ExtensionAPI) {
   const cfg = loadConfig();
-  const state: { auto: boolean; voice: string; child?: ChildProcess; lastCtx?: ExtensionContext } = {
+  const state: { auto: boolean; voice: string; child?: ChildProcess; run: number; lastCtx?: ExtensionContext } = {
     auto: cfg.auto,
     voice: cfg.voice,
+    run: 0,
   };
   const voiceLabel = () => state.voice || "server default";
   const reflectStatus = () => state.lastCtx?.ui.setStatus("dankvoice", state.auto ? `🔊 ${voiceLabel()}` : undefined);
 
   const stop = () => {
+    state.run += 1; // retire the live chain; its loop stops at the next gate
     if (state.child && state.child.exitCode === null) state.child.kill("SIGTERM");
     state.child = undefined;
   };
 
   const missingTarget = () => "dankvoice: no `api: openai-speech` provider in agent/models.yml (set DANKVOICE_URL)";
 
-  async function speak(ctx: ExtensionContext, text: string) {
-    const spoken = stripForSpeech(text, cfg.maxChars);
-    if (!spoken) return;
+  const persistSettings = (ctx: ExtensionContext) => {
+    const settings: Settings = {
+      voice: state.voice,
+      player: cfg.player?.command ?? "",
+      sink: cfg.sink,
+      maxChars: cfg.maxChars,
+      auto: state.auto,
+      thinking: cfg.thinking,
+    };
+    try {
+      writeFileSync(settingsPath(process.env), `${JSON.stringify(settings, null, 2)}\n`);
+    } catch (error) {
+      ctx.ui.notify(`dankvoice: settings not saved: ${error instanceof Error ? error.message : String(error)}`, "warning");
+    }
+  };
+
+  const setVoice = (ctx: ExtensionContext, name: string) => {
+    state.voice = name;
+    reflectStatus();
+    ctx.ui.notify(name ? `dankvoice: voice = ${name}` : "dankvoice: voice = whatever frank-tts defaults to", "info");
+    persistSettings(ctx);
+  };
+
+  const setAuto = (ctx: ExtensionContext, on: boolean) => {
+    state.auto = on;
+    if (!on) stop();
+    reflectStatus();
+    ctx.ui.notify(
+      on ? `dankvoice: auto-speak armed (${voiceLabel()} @ ${cfg.target?.speechUrl ?? "no endpoint"})` : "dankvoice: auto-speak off",
+      "info",
+    );
+    persistSettings(ctx);
+  };
+
+  /**
+   * Speak one answer as strictly sequential requests: synthesize a segment,
+   * play it to the end, only then ask for the next — audio never stacks.
+   * Like the built-in enhancer path, each unit (thinking first, then the
+   * answer) is rewritten into spoken prose by the @tiny model before
+   * segmentation; a failed rewrite falls back to the mechanical text and an
+   * empty rewrite skips its unit. Superseded chains exit at the gates.
+   */
+  async function speak(ctx: ExtensionContext, text: string, thinking = "") {
     if (!cfg.target) {
       ctx.ui.notify(missingTarget(), "error");
       return;
@@ -362,56 +922,157 @@ export default function dankvoice(pi: ExtensionAPI) {
       ctx.ui.notify("dankvoice: no audio player found (pw-play, paplay or ffplay)", "error");
       return;
     }
-    stop();
-    const dir = mkdtempSync(join(tmpdir(), "dankvoice-"));
-    try {
-      const file = await synthesize(cfg.target, speechBody(cfg.target.model, spoken, state.voice), dir, SYNTH_TIMEOUT_MS);
-      state.child = startPlayback(cfg.player, file, dir);
-    } catch (error) {
-      stop();
-      try {
-        rmSync(dir, { recursive: true, force: true });
-      } catch {
-        // ignore
-      }
-      ctx.ui.notify(`dankvoice: ${error instanceof Error ? error.message : String(error)}`, "error");
+    const target = cfg.target;
+    const player = cfg.player;
+    const units: string[] = [];
+    if (cfg.thinking && thinking.trim()) units.push(thinking.trim());
+    if (text.trim()) units.push(text.trim());
+    const segments: string[] = [];
+    let used = 0;
+    for (const unit of units) {
+      const rewritten = cfg.rewriter ? await rewriteForSpeech(cfg.rewriter, unit) : null;
+      if (rewritten === "") continue;
+      const chosen = budgetSegments(segmentsForSpeech(rewritten ?? unit), Math.max(0, cfg.maxChars - used));
+      used += chosen.reduce((chars, segment) => chars + segment.length, 0);
+      segments.push(...chosen);
     }
+    if (segments.length === 0) return;
+    stop();
+    const run = state.run;
+    for (const segment of segments) {
+      if (state.run !== run) return;
+      const dir = mkdtempSync(join(tmpdir(), "dankvoice-"));
+      try {
+        const file = await synthesize(target, speechBody(target.model, segment, state.voice), dir, SYNTH_TIMEOUT_MS);
+        if (state.run !== run) {
+          rmSync(dir, { recursive: true, force: true });
+          return;
+        }
+        await new Promise<void>((resolve) => {
+          const child = startPlayback(player, file, dir);
+          state.child = child;
+          child.once("exit", () => resolve());
+          child.once("error", () => resolve());
+        });
+      } catch (error) {
+        try {
+          rmSync(dir, { recursive: true, force: true });
+        } catch {
+          // temp files are not worth a toast
+        }
+        ctx.ui.notify(`dankvoice: ${error instanceof Error ? error.message : String(error)}`, "error");
+        return;
+      }
+    }
+  }
+
+  async function openMenu(ctx: ExtensionContext): Promise<void> {
+    const options = [
+      `Auto-speak: ${state.auto ? "on" : "off"} (toggle)`,
+      `Voice: ${voiceLabel()} (change)`,
+      `Output device: ${cfg.sink || "system default"} (change)`,
+      `Player: ${cfg.player?.command ?? "none"} (change)`,
+      `Spoken cap: ${cfg.maxChars} chars (change)`,
+      `Thinking narration: ${cfg.thinking ? "on" : "off"} (toggle)`,
+      `Humanize answers: ${cfg.rewriter ? `on (${cfg.rewriter.model})` : "off"} (toggle)`,
+      "Speak a test phrase",
+    ];
+    const choice = await ctx.ui.select("dankvoice — options", options);
+    if (!choice) return;
+    if (choice === options[0]) {
+      setAuto(ctx, !state.auto);
+      return;
+    }
+    if (choice === options[1]) {
+      if (!cfg.target) {
+        ctx.ui.notify(missingTarget(), "error");
+        return;
+      }
+      let voices: string[];
+      try {
+        voices = await fetchVoices(cfg.target);
+      } catch (error) {
+        ctx.ui.notify(`dankvoice: ${error instanceof Error ? error.message : String(error)}`, "error");
+        return;
+      }
+      const picked = await ctx.ui.select("dankvoice — voice", ["server default", ...voices]);
+      if (picked === undefined) return;
+      setVoice(ctx, picked === "server default" ? "" : picked);
+      return;
+    }
+    if (choice === options[2]) {
+      const picked = await ctx.ui.select("dankvoice — output device", ["system default", ...listSinks().map((sink) => sink.name)]);
+      if (picked === undefined) return;
+      cfg.sink = picked === "system default" ? "" : picked;
+      if (cfg.player) cfg.player = playerFor(cfg.player.command, cfg.sink);
+      ctx.ui.notify(`dankvoice: output device = ${cfg.sink || "system default"}`, "info");
+      persistSettings(ctx);
+      return;
+    }
+    if (choice === options[3]) {
+      const choices = [...new Set([cfg.player?.command, ...availablePlayers(lookupOnPath)].filter((c): c is string => Boolean(c)))];
+      if (choices.length === 0) {
+        ctx.ui.notify("dankvoice: no audio player found (pw-play, paplay or ffplay)", "error");
+        return;
+      }
+      const picked = await ctx.ui.select("dankvoice — player", choices);
+      if (picked === undefined) return;
+      cfg.player = playerFor(picked, cfg.sink);
+      ctx.ui.notify(`dankvoice: player = ${picked}`, "info");
+      persistSettings(ctx);
+      return;
+    }
+    if (choice === options[4]) {
+      const picked = await ctx.ui.select("dankvoice — spoken cap", ["300", "700", "1500", "3000"]);
+      const next = Number(picked);
+      if (!Number.isFinite(next) || next <= 0) return;
+      cfg.maxChars = next;
+      ctx.ui.notify(`dankvoice: cap = ${next} chars`, "info");
+      persistSettings(ctx);
+      return;
+    }
+    if (choice === options[5]) {
+      cfg.thinking = !cfg.thinking;
+      ctx.ui.notify(`dankvoice: thinking narration ${cfg.thinking ? "on" : "off"}`, "info");
+      persistSettings(ctx);
+      return;
+    }
+    if (choice === options[6]) {
+      cfg.rewriter = cfg.rewriter ? undefined : resolveRewriter(process.env, readModelsYml(), readOmpConfigYml());
+      ctx.ui.notify(cfg.rewriter ? `dankvoice: humanizing with ${cfg.rewriter.model}` : "dankvoice: humanization off", "info");
+      persistSettings(ctx);
+      return;
+    }
+    await speak(ctx, `dankvoice: ${voiceLabel()} through ${cfg.player?.command ?? "no player"}, checking.`);
   }
 
   pi.registerCommand("say", {
     description:
-      "Speak with the fleet's cloned voices (frank-tts): /say <text> | on | off | voice <name> | voices | stop | status",
+      "Speak with the fleet's cloned voices (frank-tts): /say (menu) | <text> | on | off | menu | voice <name> | voices | stop | status",
     handler: async (args: string, ctx) => {
       state.lastCtx = ctx;
       const action = parseSayArgs(args);
       switch (action.kind) {
+        case "menu":
+          await openMenu(ctx);
+          return;
         case "help":
           ctx.ui.notify(
-            "dankvoice: /say <text> · /say on|off (auto-speak answers) · /say voice <name> · /say voices · /say stop · /say status",
+            "dankvoice: /say (menu) · /say <text> · /say on|off (auto-speak answers) · /say voice <name> · /say voices · /say stop · /say status",
             "info",
           );
           return;
         case "on":
-          state.auto = true;
-          reflectStatus();
-          ctx.ui.notify(`dankvoice: auto-speak armed (${voiceLabel()} @ ${cfg.target?.speechUrl ?? "no endpoint"})`, "info");
+          setAuto(ctx, true);
           return;
         case "off":
-          state.auto = false;
-          stop();
-          reflectStatus();
-          ctx.ui.notify("dankvoice: auto-speak off", "info");
+          setAuto(ctx, false);
           return;
         case "stop":
           stop();
           return;
         case "voice":
-          state.voice = action.name;
-          reflectStatus();
-          ctx.ui.notify(
-            action.name ? `dankvoice: voice = ${action.name}` : "dankvoice: voice = whatever frank-tts defaults to",
-            "info",
-          );
+          setVoice(ctx, action.name);
           return;
         case "voices": {
           if (!cfg.target) {
@@ -435,18 +1096,18 @@ export default function dankvoice(pi: ExtensionAPI) {
         }
         case "status":
           ctx.ui.notify(
-            `dankvoice: voice ${voiceLabel()} · ${cfg.target?.speechUrl ?? "NO ENDPOINT"} · player ${cfg.player ? cfg.player.command : "NONE"} · auto-speak ${state.auto ? "on" : "off"} · cap ${cfg.maxChars} chars`,
+            `dankvoice: voice ${voiceLabel()} · ${cfg.target?.speechUrl ?? "NO ENDPOINT"} · player ${cfg.player ? cfg.player.command : "NONE"} · device ${cfg.sink || "system default"} · auto-speak ${state.auto ? "on" : "off"} · cap ${cfg.maxChars} chars · thinking ${cfg.thinking ? "on" : "off"} · humanize ${cfg.rewriter ? cfg.rewriter.model : "off"}`,
             "info",
           );
           return;
         case "speak":
-          await speak(ctx, action.text);
+          void speak(ctx, action.text).catch((error) => ctx.ui.notify(`dankvoice: ${error instanceof Error ? error.message : String(error)}`, "error"));
           return;
       }
     },
   });
 
-  pi.on("agent_end", async (event, ctx) => {
+  pi.on("agent_end", (event, ctx) => {
     state.lastCtx = ctx;
     if (!state.auto) return;
     // `willContinue` means the harness already scheduled a continuation
@@ -454,7 +1115,12 @@ export default function dankvoice(pi: ExtensionAPI) {
     // settle, so narrating it would speak an answer the model is about to redo.
     if (event.willContinue) return;
     if (ctx.mode === "print" || ctx.mode === "json") return;
-    await speak(ctx, lastAssistantText(event.messages));
+    // Fire-and-forget: a multi-segment utterance legitimately outlives the
+    // harness's handler timeout, and blocking here would only produce a
+    // spurious "handler timed out" toast while playback continues anyway.
+    void speak(ctx, lastAssistantText(event.messages), lastAssistantThinking(event.messages)).catch((error) =>
+      ctx.ui.notify(`dankvoice: ${error instanceof Error ? error.message : String(error)}`, "error"),
+    );
   });
 
   pi.on("session_shutdown", () => {

agent/managed-skills/bounding-subagent-runtime/SKILL.md

@@ -1,82 +1,32 @@
 ---
 name: bounding-subagent-runtime
-description: "Cap and diagnose runaway or hung omp subagents: task.maxRuntimeMs, softRequestBudget, agentIdleTtlMs, per-agent model overrides, the cfg save-approval quirk, and the finished-work-but-never-exits nested-fanout deadlock."
+description: "Cap and diagnose runaway or hung omp subagents: task.maxRuntimeMs, softRequestBudget, maxConcurrency, maxRecursionDepth (the 0-trap), per-provider in-flight ceilings."
 ---
 
-# Bounding subagent runtime (omp harness)
+## When
+Subagent runs long, fan out too wide, or you must (re)declare the fleet's subagent limit policy. Also when asked "what limits can subagents have".
 
-Use when a `task`/`eval` subagent runs absurdly long (tens of minutes on a small job),
-when it **finishes its deliverable and still never exits**, when you want a hard cap
-before spawning, or when you must tell a busy agent to stop.
+## Read first
+- Repo KB page `omp-docs/pi/subagent-limits.md` — the live policy of record (values + rationale + deliberately-not-set list). Do not re-derive; read it.
+- `cfg://task` and `cfg://providers/maxInFlightRequests` — effective values, not the stock defaults.
 
-## The knobs (`cfg://task/...`)
-
-| Key | Default | Meaning |
+## The knobs (cfg:// writes; `/save` suffix persists)
+| Knob | Effect | Gotcha |
 |---|---|---|
-| `maxRuntimeMs` | `0` (unlimited) | **Hard wall-clock limit per subagent (ms).** The only true time cap. |
-| `softRequestBudget` | `200` | Soft cap on assistant **requests** per run; force-stop at **1.5×**. |
-| `softRequestBudgetNotice` | `true` | Injects one "wrap up" steering message at the soft mark. |
-| `agentIdleTtlMs` | `420000` | Parks an **idle** agent after 7 min. Useless against a busy-but-slow agent. |
-| `completionProbe` | `true` | Asks a working subagent to self-report completeness at 2/5/10 requests. |
-| `maxConcurrency` | `32` (often lowered) | Concurrent subagents; sizes GPU/endpoint load. |
-| `agentModelOverrides` | `{}` | Per-agent model routing, e.g. `{"scout":"provider/model-id"}`. |
-| `tools.maxTimeout` | `0` | Caps the timeout the agent may set on a **tool**, not on the agent. |
+| `task.maxConcurrency` | session semaphore; excess queues | stock default 32 — "2" is not a default |
+| `task.maxRecursionDepth` | nesting cap | predicate is `depth < n`; **1 = children cannot spawn (correct for "no nesting"); 0 disables spawning for the MAIN session too — never reach for 0**; negative = unlimited |
+| `task.maxRuntimeMs` | hard wall clock, overrun = `aborted` | default 0 = off |
+| `task.softRequestBudget` | wrap-up notice at N, force-yield at 1.5N | scout/sonic may impose lower built-ins |
+| `task.agentIdleTtlMs` | park idle finished agents | `<= 0` disables parking |
+| `providers.maxInFlightRequests` | concurrent LLM requests per provider id | keys are the quoted block labels from `agent/models.yml` (e.g. `"llamacpp (dankclaw-3090)"`), flat map lookup; absent/≤0 = unlimited; scope = this machine's config root only — other workstations and non-omp clients are NOT counted; excess queues |
+| `tools.maxTimeout` | ceiling on any agent-requested tool timeout | kills `timeout: 0` runaways |
+| `task.isolation.enabled` | makes per-spawn `isolated` flag available | opt-in per spawn, not a blanket sandbox; needs git repo; off in plan mode |
+| `task.completionProbe` | periodic completeness side-requests | token sink on shared endpoints |
 
-Reasonable fleet defaults: `maxRuntimeMs = 900000` (15 min), leave the request budget
-alone unless the failure mode is a looping search agent (then lower it — request count,
-not the clock, is what bounds a 200-call research loop).
+## Verify semantics from the bundle, not the label
+Setting descriptions lie by omission. `~/.bun/install/global/node_modules/@oh-my-pi/pi-coding-agent/dist/cli.js`:
+- depth gate: `grep -o '.\{130\}xC\b.\{130\}'` finds `U_(n,s)`; `grep -o 'function U_(e,t){.\{0,60\}'` → `e<0||t<e` (n=cap, s=taskDepth).
+- in-flight lookup: `grep -o '.\{150\}axInFlightRequests.\{200\}'` → flat `[providerId]` lookup, non-number/≤0 = unlimited, clamp ≥1.
 
-**Verify the cap instead of assuming it.** `read cfg://task/maxRuntimeMs` at the start of
-any session that plans to rely on it. A session-only write is gone next session, and an
-unanswered `/save` approval means the value never persisted — so the fleet can run hours
-with the cap at `0` while everyone believes it is 15 minutes. Nothing else bounds a child.
-
-## Three failure modes, three different fixes
-
-- **Runaway loop** (hundreds of sequential tool calls): lower `softRequestBudget`; the
-  clock cap only truncates it late.
-- **Slow model, legitimate work** (dozens of calls on a low-t/s endpoint = an hour of wall
-  clock): fix routing, not the clock — check `cfg://task/agentModelOverrides` and move the
-  agent class onto a faster/shared endpoint. Watch the slot math: subagent endpoints share
-  KV slots with the interactive session, so `maxConcurrency + 1` sequences on an N-slot
-  endpoint degrades everyone's decode.
-- **Finished work, no exit (nested fan-out deadlock).** A subagent that spawned **its own**
-  children blocks on their barriers. If those children die out-of-band (killed, starved, or
-  waiting on an answer whose subject changed mid-flight), the parent waits forever with its
-  deliverable already written. No instruction text prevents this — not the repo AGENTS.md,
-  not the task prompt; only the clock cap or a kill ends it.
-
-  Signature: parent and children report **the same elapsed wall clock**, and the parent's
-  on-disk output is complete. Census (agent registry + `ps`): group by elapsed time, look
-  for identical `etimes` across a parent/child set, then judge the **artifacts** instead of
-  the process — balanced code fences, no truncated tables, trailing newline, front matter or
-  the page's H1+status convention intact. Complete artifacts + still "running" = deadlock,
-  not work.
-
-  Repair: kill the **subtree** and confirm the children leave (a killed parent does not
-  cancel children that came from a separate `tasks[]` batch). Then **flatten**: the
-  top-level session does the fan-out and children never spawn children. Re-run only the
-  slices whose artifacts are missing — integrate from what is already on disk.
-
-## cfg write mechanics (bit me once)
-
-- `write cfg://task/maxRuntimeMs 900000` → applies **session-only**, replies "NOT saved".
-- Persist = `write cfg://task/maxRuntimeMs/save` with the same content, and that write
-  needs its **own interactive approval**. If the approval prompt is not answered in time
-  the value stays session-only and the tool says *NEVER retry unless the user asks* —
-  report it in the reply instead.
-- Revert a session value by writing the old value back (e.g. `0`).
-
-## Order of operations when one is already running late
-
-1. `read proc://` → registry of running agents; `read proc://<id>` for status/output.
-2. `read agent://<id>` → live progress ("Latest assistant text") — often shows it is
-   mid-batch, not hung. If the latest text is already a wrap-up/summary, treat it as
-   **done** and collect its output instead of waiting for an exit event.
-3. `write agent://<id>` an explicit wrap-up order: stop searching, answer what is
-   sourced, mark the rest `NOT VERIFIED — needs <what>`, cap the report length.
-4. Only then `write proc://<id>/kill` — the cap does **not** retroactively bind agents
-   spawned before the change, so a long runner must be ordered or killed, never just
-   re-configured. After any kill, re-read the registry: unreaped children and orphaned
-   host-side processes (a scratch server holding GPU/RAM) are the expensive residue, and
-   `pkill -f` patterns that also match production services are forbidden — kill by pid.
+## Landing a policy change
+cfg `/save` writes persist immediately; `agent/config.yml` is machine-live tier (heartbeat commits it). Human-facing policy changes land as an `omp-docs` page update via branch → PR (docs tier), `git worktree add` from `origin/main` if the current checkout is branch-locked by workspace policy.

agent/memories/sharpshooter/omp-32c11n2wnglul/state.json

@@ -1,6 +1,6 @@
 {
 	"v": 1,
-	"lastConsolidatedAt": 1792381479415,
+	"lastConsolidatedAt": 1792395788647,
 	"lastResult": {
 		"at": 1791599865775,
 		"sessions": 1,

agent/memories/sharpshooter/tmp-1tpw73408x3e/state.json

@@ -1,4 +1,4 @@
 {
 	"v": 1,
-	"lastConsolidatedAt": 1792392829703
+	"lastConsolidatedAt": 1792396789735
 }

agent/memories/sharpshooter/wt-20261017-181502-6d1fc1f-1oq2osmoyw03k/architecture.md

@@ -1,3 +1,20 @@
+## Subagent routing (harness law, 2026-10-18)
+
+**No subagent class may route to `frank-archon`** — the operator session lives there and
+the endpoint runs 3 sequences (`--max-num-seqs 3`), so `task.maxConcurrency + 1` children
+eat the parent's own KV slots. Violation cost: three doc children + parent = 4 sequences
+on 3 slots, children ran 2 h 27 m / 2 h 45 m, the interactive session starved.
+Routing and bounds live in `~/.omp/agent/config.yml` (`task:` block — the owner; verified
+with `read cfg://task/agentModelOverrides`): heavy writers (`task`, `scout`) → frank-dragoon
+(2 slots, idle, slow card ⇒ slice work smaller), bursty readers (`reviewer`, `sonic`,
+`designer`, `librarian`) → frank-zealot (2 slots, also the compaction hub and every
+tiny/smol/plan/commit/memory role). Caps: `maxConcurrency: 2`, `maxRuntimeMs: 1800000`,
+`softRequestBudget: 120`, `maxRecursionDepth: 1` (children never spawn children).
+`cfg://.../save` approvals auto-deny after a hardcoded 10 s dialog — after Frank has
+approved a change, edit `config.yml` directly and validate the YAML; a running session
+picks global-config edits up at once. Full procedure: managed skill
+`bounding-subagent-runtime`.
+
 ## Documentation law (2026-10-18)
 
 **One fact, one owner.** The `serverAppConfigs` repo owns every fleet fact;
@@ -18,8 +35,8 @@ Swap the model, don't instrument the swap: edit the trio
 probe → done. **No benchmarks, A/Bs, ladders, or acceptance tables unless
 Frank explicitly asks for data.** If a payload misbehaves, revert the commit and
 redeploy. When a bench IS ordered, the GPU-safety procedure (stop + mask, no
-scratch ports, never `systemctl disable`) is `AGENTS.md` → *Model swaps* and
-*Deployment completeness*.
+scratch ports, never `systemctl disable`) is `AGENTS.md` → *Model swaps*
+and *Deployment completeness*. 
 
 ## Where facts live
 

agent/memories/sharpshooter/wt-20261017-181502-6d1fc1f-1oq2osmoyw03k/product.md

@@ -3,25 +3,22 @@
 - Custom voice source material must exclude the first 10 seconds of any audio sample file used to generate voices; this is a project-wide invariant across all source files, not a per-file exception.
 - Voice selector valid options: af_heart, af_alloy, af_aoede, af_bella, af_jessica, af_kore, af_nicole, af_nova, af_river, af_sarah, af_sky, am_adam, am_echo, am_eric, am_fenrir, am_liam, am_michael, am_onyx, am_puck, am_santa, bf_emma, bf_isabella, bm_george, bm_lewis, bf_alice, bf_lily, bm_daniel, bm_fable.
 
-## Podcast
-
-- Podcast functionality is sourced from open-notebook; this is a project decision established before this document was created.
-
-## Audio Input Policy
-
-- There is no audio input anywhere in the fleet: that capability left with the FLM NPU endpoint on 2026-09-11, returned 2026-09-17 with the Gemma4 projector's audio encoder, and left again with the Gemma4 retirement on 2026-10-01.
-- This policy mirrors the architecture-level decision documented in `architecture.md` regarding the fleet's input capabilities.
-
 ## Input Capabilities
 
 All endpoints accept text input with optional image input: `frank-dragoon` returned to vision with the flash-server v2 swap, while `frank-archon` takes them natively and `frank-zealot` through its BF16 mmproj.
 
 There is no audio input anywhere in the fleet: that capability left with the FLM NPU endpoint on 2026-09-11, returned 2026-09-17 with the Gemma4 projector's audio encoder, and left again with the Gemma4 retirement on 2026-10-01.
 
-This rule exists because the Gemma4 projector was retired in 2026-10-01, removing the audio input channel from the fleet. The 2026-09-11 to 2026-10-01 period contained the Gemma4 projector which enabled audio, but that endpoint is now retired.
-
 ## Voice Usage Notes
 
 - Voice generator output does not affect token counting or inference costs — all voices are rendered post-inference.
 - Voice quality varies by source sample: exclude the first 10 seconds of any audio sample file used to generate voices; this is a project-wide invariant across all source files, not a per-file exception.
-- Voice selection is client-side only; the server does not generate or return audio data.`, 
+- Voice selection is client-side only; the server does not generate or return audio data.
+
+## Load Management
+
+- Subagent settings must prevent frank-archon from being overwhelmed; concurrent request handling prioritizes endpoint stability over throughput.
+
+## Podcast
+
+- Podcast functionality is sourced from open-notebook; this is a project decision established before this document was created.

agent/memories/sharpshooter/wt-20261017-181502-6d1fc1f-1oq2osmoyw03k/state.json

@@ -1,8 +1,8 @@
 {
 	"v": 1,
-	"lastConsolidatedAt": 1792392891753,
+	"lastConsolidatedAt": 1792396748497,
 	"lastResult": {
-		"at": 1792379504653,
+		"at": 1792393886960,
 		"sessions": 1,
 		"deltas": 1,
 		"model": "frank-zealot"

agent/memories/sharpshooter/wt-20261018-204442-7c8a99d-2u1dm4kttza8t/product.md

@@ -0,0 +1,3 @@
+# Output behavior
+
+Output must not be shaken off during presentation. Output remains anchored until explicitly confirmed.

agent/memories/sharpshooter/wt-20261018-204442-7c8a99d-2u1dm4kttza8t/state.json

@@ -1,4 +1,10 @@
 {
 	"v": 1,
-	"lastConsolidatedAt": 1792392882348
+	"lastConsolidatedAt": 1792396722373,
+	"lastResult": {
+		"at": 1792396071933,
+		"sessions": 1,
+		"deltas": 1,
+		"model": "frank-zealot"
+	}
 }

agent/memories/sharpshooter/wt-20261018-204501-6d1fc1f-3fct0wvty8jvu/state.json

@@ -1,6 +1,6 @@
 {
 	"v": 1,
-	"lastConsolidatedAt": 1792392961865,
+	"lastConsolidatedAt": 1792393561866,
 	"lastResult": {
 		"at": 1792381870670,
 		"sessions": 1,

agent/memories/sharpshooter/wt-20261019-001426-7c8a99d-3m7k3cagiow5n/queue/01a15302-f12d-77d8-9a33-be31a1a05528/00mvexe0pf-6ozb.json

@@ -0,0 +1 @@
+{"v":1,"kind":"product_decision","statement":"Limit on zealot set to 3, dragoon set to 2, archon set to 3","source":"explicit_user","evidence":"make the limit on zealot 3, dragoon 2, archon 3","friction":{"corrective":false,"regression":false,"subtle":false},"sessionId":"01a15302-f12d-77d8-9a33-be31a1a05528","ts":1792394470419}

agent/memories/sharpshooter/wt-20261019-001426-7c8a99d-3m7k3cagiow5n/state.json

@@ -0,0 +1,4 @@
+{
+	"v": 1,
+	"lastConsolidatedAt": 1792394427081
+}

agent/memories/sharpshooter/wt-20261019-004309-7c8a99d-u0xusjwyvzai/state.json

@@ -0,0 +1,4 @@
+{
+	"v": 1,
+	"lastConsolidatedAt": 1792396749853
+}

dotfiles/noctalia/.local/state/noctalia/settings.toml

@@ -275,13 +275,13 @@ transition = [ "fade" ]
     interval_seconds = 1200
 
     [wallpaper.default]
-    path = "/home/blank/.local/share/noctalia/wallpapers/wallhaven_4x17w3.jpg"
+    path = "/home/blank/.local/share/noctalia/wallpapers/wallhaven_nzz3gy.jpg"
 
     [wallpaper.last]
-    path = "/home/blank/.local/share/noctalia/wallpapers/wallhaven_4x17w3.jpg"
+    path = "/home/blank/.local/share/noctalia/wallpapers/wallhaven_nzz3gy.jpg"
 
     [wallpaper.monitors.DP-1]
-    path = "/home/blank/.local/share/noctalia/wallpapers/wallhaven_4x17w3.jpg"
+    path = "/home/blank/.local/share/noctalia/wallpapers/wallhaven_nzz3gy.jpg"
 
     [wallpaper.monitors.DP-2]
     path = "/home/blank/.local/share/noctalia/wallpapers/wallhaven_oxex69.jpg"
@@ -290,7 +290,7 @@ transition = [ "fade" ]
     path = "/home/blank/.local/share/noctalia/wallpapers/wallhaven_v9e2w3.jpg"
 
     [wallpaper.monitors.eDP-1]
-    path = "/home/blank/.local/share/noctalia/wallpapers/wallhaven_4x17w3.jpg"
+    path = "/home/blank/.local/share/noctalia/wallpapers/wallhaven_nzz3gy.jpg"
 
 [weather]
 unit = "imperial"

omp-docs/core/dankvoice.md

@@ -9,17 +9,19 @@ tags:
   - voice
   - frank-tts
   - dankclaw
-verified: 2026-10-18
-timestamp: 2026-10-18
+verified: 2026-10-19
+timestamp: 2026-10-19
 ---
 
 # dankvoice
 
 `agent/extensions/dankvoice.ts` adds cloned-voice speech to the interactive
-harness: `/say <text>` speaks through the speakers and `/say on` narrates the
-assistant's final answer after every turn, both served by
-**frank-tts** (Dankclaw :7772, the `speech` model role in
-[agent/models.yml](/core/index.md)).
+harness: `/say` opens the options menu, `/say <text>` speaks through the
+speakers and `/say on` narrates the assistant's thinking and final answer
+after every turn. The words are first rewritten into natural spoken prose by
+the **@tiny** role model (exactly what the built-in vocalizer's
+`speech.enhanced` does), and everything is served by **frank-tts**
+(Dankclaw :7772, the `speech` model role in [agent/models.yml](/core/index.md)).
 
 ## Why it exists
 
@@ -39,17 +41,58 @@ local speech and the cloned voices carry the real narration.
 
 | Command | Effect |
 |---|---|
+| `/say` \| `/say menu` | options menu (`ctx.ui.select`): auto-speak toggle, voice, output device, player, spoken cap, thinking-narration toggle, humanize toggle, test phrase |
 | `/say <text>` | synthesize once and play; interrupts whatever is playing |
-| `/say on` \| `/say off` | arm/disarm auto-speak of the assistant's final text at `agent_end`; the footer shows `🔊 <voice>` while armed |
+| `/say on` \| `/say off` | arm/disarm auto-speak of the thinking blocks and final text at `agent_end`; the footer shows `🔊 <voice>` while armed |
 | `/say voice <name>` | pick a voice; an empty name defers to the server's own default |
 | `/say voices` | list what frank-tts reports |
 | `/say stop` | cut the current utterance |
-| `/say status` \| `/say help` | endpoint, voice, player, cap |
+| `/say status` \| `/say help` | endpoint, voice, player, device, cap, thinking, humanize |
+
+Menu choices persist to `~/.omp/cache/dankvoice.json` (runtime state, never
+tracked); the environment overrides the file, the file overrides defaults.
+"Speak a test phrase" synthesizes a short line through the current voice,
+player and device — the audible proof whenever a session "has say on" but
+stays silent.
 
 Auto-speak is skipped in `print`/`json` mode so benches and pipelines stay
-silent. Markdown is stripped before synthesis — fenced code is dropped, links
-collapse to their label, bullets and table pipes disappear — and the utterance
-is truncated at a sentence boundary inside the cap.
+silent. Each spoken unit — thinking first when narrated, then the answer —
+is first rewritten into spoken prose by the @tiny model; when the enhancer
+(below) declines, the text goes to segmentation as-is, where fenced code and
+tables vanish, links collapse to their label and bullets and table pipes
+disappear. A long answer is never one monolithic synthesis request; see
+[Segmentation](#segmentation) for how utterances are cut and played.
+
+## Segmentation
+
+Utterances are cut by a state machine ported from the harness's own
+speak-aloud vocalizer (the minified `fHe` class in the agent bundle's
+`dist/cli.js`), kept deliberately in lockstep with the stock path:
+
+* A strong boundary is a run of `[.!?…]` followed by whitespace — so
+  `!!!!!` and `...` are one boundary, never a pile of cuts.
+* Boundaries after `e.g.`, `etc.`, `Mr.` and friends, or inside inline
+  code, don't count; decimals (`3.5`, `1,5`) survive because a boundary
+  requires trailing whitespace.
+* Line starts are classified: table rows and fenced code are swallowed
+  whole (their content is never read aloud), heading and bullet markers
+  vanish, ordered-list markers speak their number ("1, …"), quote markers
+  strip.
+* Per segment, markdown is cleaned: links collapse to their label, URLs
+  to their host, deep paths to their last segment; emphasis markup dies.
+* Segments aim for one sentence; run-ons break at commas or word wraps
+  before 280 characters.
+
+Playback is strictly sequential: segment N+1 is synthesized only after
+segment N's player process exited, so audio never stacks. The chain runs
+detached from `agent_end` (fire-and-forget) and may legitimately outlive
+the harness's 30 s handler timeout — the old awaiting design produced a
+"handler timed out after 30000ms" toast on long answers. A new utterance,
+`/say stop`, or disarming auto-speak retires the live chain through a run
+token: the current player is killed and pending segments are skipped.
+`DANKVOICE_MAX` is a per-answer character budget: the leading whole
+segments that fit are spoken, the rest are dropped — the thinking unit
+draws on it before the answer does.
 
 ## Endpoint resolution
 
@@ -63,18 +106,50 @@ resolves, so no host address and no credential appears in this file. From
 |---|---|---|
 | `DANKVOICE_URL` | models.yml | speech base URL (`…/v1`), bypasses the YAML read |
 | `DANKVOICE_MODEL` | the id in that block (`frank-tts`) | model named in the request body |
+| `DANKVOICE_CONFIG` | `~/.omp/cache/dankvoice.json` | persisted menu choices (voice, player, sink, cap, auto) |
 | `DANKVOICE_VOICE` | empty (server default) | default voice name |
 | `DANKVOICE_AUTO=1` | off | start with auto-speak armed |
-| `DANKVOICE_MAX` | `700` | spoken characters per utterance |
+| `DANKVOICE_MAX` | `700` | spoken-character budget per answer (whole leading segments, shared by the thinking and answer units) |
 | `DANKVOICE_PLAYER` | detect | player binary, skipping detection |
 | `DANKVOICE_MODELS` | `~/.omp/agent/models.yml` | YAML to read the endpoint from |
+| `DANKVOICE_REWRITE_URL` | the @tiny role | chat base URL (`…/v1`) for the enhancer, bypassing role resolution |
+| `DANKVOICE_REWRITE_MODEL` | the @tiny role's model | chat model named in the rewrite body |
+| `DANKVOICE_REWRITE=0` | enhancer on when a route resolves | speak mechanical text |
+| `DANKVOICE_THINKING` | settings file, default on | `0`/`1` force thinking narration off/on |
+| `DANKVOICE_OMPCONFIG` | `~/.omp/agent/config.yml` | YAML to read the `modelRoles.tiny` role from |
 
-Player detection order is `pw-play` → `paplay` → `ffplay -nodisp -autoexit`
-(PipeWire first: the workstations run WirePlumber). Each utterance is written
-to `$TMPDIR/dankvoice-*/say-<ms>.wav` and the file plus its directory are
-removed when the player process exits.
+Player detection order is `pw-play` → `paplay` → `ffplay` (PipeWire first:
+the workstations run WirePlumber). A saved output device routes the player
+explicitly — `pw-play --target <sink>` or `paplay --device=<sink>`; ffplay
+has no device concept and ignores it — while an unset device leaves the
+player's own default. Sinks are enumerated with `wpctl list audio sinks`;
+if wireplumber is unreachable the menu offers only "system default". Each
+utterance is written to `$TMPDIR/dankvoice-*/say-<ms>.wav` and the file plus
+its directory are removed when the player process exits.
 
-## Verified facts (2026-10-18)
+## Humanized narration (the @tiny role)
+
+The built-in vocalizer never reads answers aloud verbatim when
+`speech.enhanced` is on: a small model rewrites each utterance into words a
+person would say. `dankvoice` ports that enhancer — same system prompt
+(verbatim from the bundle), same budgets (1536 max tokens, 6 s timeout,
+4000-char input window), same contract (a failed rewrite falls back to the
+mechanically cleaned text; an empty rewrite skips the unit as "nothing to
+say"). The target model is the harness's `@tiny` role, resolved the way the
+harness resolves it: `modelRoles.tiny` in `agent/config.yml` reads
+`provider (host)/model:effort`; the provider block in `agent/models.yml`
+supplies the `baseUrl`; the request is `POST {base}/chat/completions`. On
+this fleet that is **frank-zealot** (`llamacpp (dankclaw-3090)`,
+`:7771`) — the 6 s abort keeps a slow tiny model off the critical path.
+
+Thinking blocks ride the same path: at `agent_end` the last assistant
+message's true `thinking` parts are narrated (rewritten first, as their own
+unit, before the answer), while blocks flagged `narration: true` — which
+providers use to address the user through the thinking channel — count as
+answer text. The menu, `/say status`, and `DANKVOICE_THINKING` toggle it;
+the choice persists in the settings file.
+
+## Verified facts (2026-10-18, extended 2026-10-19)
 
 * **Extension discovery** — any `.ts` file in `~/.omp/agent/extensions/` loads
   at startup: a probe file there printed `loaded 2026-10-19T02:36:55Z` from a
@@ -101,6 +176,26 @@ removed when the player process exits.
   (1.20 s WAV), and an unknown voice name answers with the roster instead of
   speaking. An `agent_end` carrying `willContinue` (the harness scheduled an
   auto-retry) is never narrated — only the terminal settle speaks.
+* **Menu** — verified via driver scenarios 2026-10-18 (Refs #23): the menu
+  lists auto-speak/voice/device/player/cap/test-phrase; a voice chosen in the
+  menu is used by later utterances and survives a fresh load via the settings
+  file; a device chosen in the menu appears in the player invocation as
+  `--target <sink>`.
+* **Segmented playback** — driver scenarios 2026-10-19: "First stop here.
+  Second one goes!!! Really? Third ... done" produced exactly 3 sequential
+  requests and 3 plays; "Model 3.5 ships today. Version 1,5 too." produced
+  exactly 2 (decimals never split); the markdown fixture's fenced code and
+  URL never reached the fixture endpoint; with `DANKVOICE_MAX=60` only the
+  leading segment was spoken.
+* **Humanize + thinking (2026-10-19, Refs #27)** — driver scenarios cover the
+  rewrite round-trip (fixture chat route), the 404 → mechanical fallback,
+  thinking narrated as its own unit before the answer, `DANKVOICE_THINKING=0`
+  silencing it, and `narration: true` spoken once as the answer; the endpoint
+  status line shows the resolved humanize model from the @tiny role. Live
+  smoke against the real services: a markdown `/say` became 3 sequential
+  prose segments (mechanical cleanup of that text yields at most 2), and a
+  thinking+answer `agent_end` played thinking first. 22 green in
+  `tests/dankvoice.bats`.
 
 ## Limits
 
@@ -108,8 +203,13 @@ removed when the player process exits.
   a compile-time enum in the harness bundle, and no setting routes it to
   frank-tts. Cloned voices live only on this file-producing/speaking path and
   on the `tts` tool. See [Handy](/tools/handy.md) for the STT half.
-* `agent_end` narrates the final message text only — mid-turn tool chatter is
-  never spoken.
+* `agent_end` narrates thinking blocks and the final message only — mid-turn
+  tool chatter is never spoken. Thinking covers the last assistant message;
+  tool-loop gaps in earlier messages are skipped by design.
+* Humanized narration is serial: audio starts only after the @tiny rewrite
+  returns (≤ 6 s per unit, then frank-tts synthesis). Say `/say off`-grade
+  latency is the price of prose; `DANKVOICE_REWRITE=0` trades it back for
+  instant mechanical speech.
 * frank-tts is LAN-only: no Caddy vhost fronts :7772, so this extension is
   usable on the home network.
 * Deployment, packs and the voice registry are owned by the
@@ -124,7 +224,14 @@ removed when the player process exits.
 * `omp --help` (extension discovery flags, `say`, `plugin`) and
   `@earendil-works/pi-coding-agent` `dist/core/extensions/types.d.ts`
   (`ExtensionAPI.on/registerCommand`, `ExtensionMode = "tui" | "rpc" | "json" | "print"`).
+* The stock speak-aloud implementation — vocalizer pipeline `I6o`, utterance
+  segmenter `fHe` and speech enhancer `Qyr` with its prompt `T6o` — in
+  `@oh-my-pi/pi-coding-agent`'s `dist/cli.js`, read off the minified bundle
+  2026-10-19: dankvoice ports the mechanical and the enhanced halves, so an
+  LLM rewrite does happen before frank-tts.
 * Gitea issue #21 on `dank/command-center` (the work item this page ships with).
+* Gitea issue #23 on `dank/command-center` (the options-menu work item).
+* Gitea issue #27 on `dank/command-center` (the enhanced-narration work item).
 
 ## See Also
 

omp-docs/log.md

@@ -11,6 +11,60 @@ timestamp: 2026-08-28
 
 # Bundle Change Log
 
+
+## 2026-10-19 — dankvoice: humanized narration and thinking aloud (Refs #27)
+
+dankvoice answered in verbatim markdown-cleaned text while the built-in
+vocalizer speaks humanized prose — the bundle's `Qyr` enhancer (`speech.enhanced`,
+system prompt `T6o`) now lives in the extension: every spoken unit is rewritten
+by the @tiny role (resolved from `modelRoles.tiny` in config.yml + the provider
+baseUrl in models.yml → frank-zealot :7771) under the same budgets (1536 tokens,
+6 s, 4000-char window), with mechanical cleanup as the fallback and an empty
+rewrite skipping the unit. Auto-speak also narrates the last assistant message's
+thinking blocks before the answer (`narration: true` blocks count as answer
+text); both are menu-toggleable and persisted (`thinking` in the settings file),
+env levers `DANKVOICE_REWRITE_URL/MODEL/REWRITE=0/THINKING/OMPCONFIG`. The
+fixture grows a chat route with echo/prose modes; 7 new scenarios, 22 green.
+Live smoke: a markdown answer became 3 sequential prose segments through real
+zealot + frank-tts, thinking played before its answer.
+
+## 2026-10-19 — dankvoice: segmented sequential speech, copied from the bundle's vocalizer
+
+Silent-answer incident follow-up: `agent_end` awaiting one whole-answer
+synthesis hit the harness's 30 s extension-handler timeout
+(`handler timed out after 30000ms`) on long answers, and the single frank-tts
+request for the entire answer was the stall. Instead of inventing a heuristic,
+ported the utterance-splitting state machine out of the built-in speak-aloud
+vocalizer (minified class `fHe` in `dist/cli.js`): strong boundaries are
+`[.!?…]+` runs followed by whitespace — `!!!!!`/`...` stay one cut — with
+abbreviation and inline-code guards; `3.5`/`1,5` survive because a boundary
+needs trailing whitespace; fenced code and tables swallow whole; headings and
+bullets vanish, ordered markers speak their number ("1, "); links collapse to
+labels, URLs to hosts, paths to basenames; run-ons break at commas or word
+wraps before 280 chars. `speak()` now synthesizes segment N+1 only after
+segment N's player exited — strictly sequential, audio never stacks — and
+`agent_end`/`/say <text>` fire-and-forget a chain retired through a run token
+by `/say stop`, disarming, or a newer utterance. `DANKVOICE_MAX` became a
+per-answer budget of whole leading segments; the hand-rolled `stripForSpeech`
+truncator is gone. No LLM "humanization" rewrite happens (the bundle's
+`speech.enhanced` rewrites with the tiny/smol model; dankvoice speaks
+mechanically cleaned text through the cloned voices). Tests: two new driver
+scenarios plus two rewrites, 17/17 in `tests/dankvoice.bats`.
+
+## 2026-10-18 — dankvoice: `/say` options menu with output-device routing (Refs #23)
+
+Added `/say menu` (bare `/say` now opens it; `help` keeps the text listing), the
+pi-peon-style settings surface: auto-speak toggle, voice picked from the frank-tts
+roster, **output device** (PipeWire sinks via `wpctl` → `pw-play --target` /
+`paplay --device=`), player binary, spoken-char cap, and a "speak a test phrase"
+utterance that proves the audio path audibly. Choices persist to
+`~/.omp/cache/dankvoice.json` (`DANKVOICE_CONFIG`); env now wins over that file,
+which wins over defaults — answering the silent-session incident where
+PipeWire's remembered Bluetooth route played synthesized answers into nowhere.
+`detectPlayer` becomes `availablePlayers` + `playerFor`. Tests gain a fixture
+`wpctl`, a scripted `select` in the driver, and three menu scenarios: 15 green
+in `tests/dankvoice.bats`, 51 green suite-wide.
+
 ## 2026-10-18 — dankvoice: the cloned voices finally speak in the TUI (Refs #21)
 
 Nothing in the harness could pronounce them: the TUI vocalizer and `omp say` gate `--voice` against the bundled Kokoro enum (`omp say --voice narrator1 "test"` → `error: Expected --voice to be one of: af_heart, …`), so frank-tts was reachable only as a file. Added `agent/extensions/dankvoice.ts` — `/say <text>`, `/say on|off` (narrate the final answer at `agent_end`), `/say voice|voices|stop|status` — synthesizing over the OpenAI speech route and playing through `pw-play`/`paplay`/`ffplay`, with the endpoint read from the `api: openai-speech` row of `agent/models.yml` (no host, no secret in the file). Verified live in a real TUI session on the workstation: roster `GET /v1/audio/voices` → `narrator1, narrator2, narrator3`; `/say` produced 99,884 B / 2.08 s and 69,164 B / 1.44 s WAVs handed to the player, temp dirs cleaned afterwards; `/say on` armed the footer (`🔊 narrator1`) and each following finished turn spoke its answer (1.20 s WAV), an unknown voice name answered with the roster instead of speaking, and an `agent_end` carrying `willContinue` (the harness scheduled an auto-retry) is deliberately never narrated. Extension discovery itself proved with a probe file in `~/.omp/agent/extensions/` loaded by a fresh `omp -p` run. Reference page [dankvoice](core/dankvoice.md), registered in the core index; black-box seam `tests/dankvoice.bats` (12 tests).

tests/dankvoice.bats

@@ -3,15 +3,19 @@
 #
 # The extension's contract, from outside: it registers a `say` command and an
 # `agent_end` listener; what is observable is (a) the JSON it POSTs to the
-# speech route, (b) the file it hands to the player, and (c) the notification
-# text it prints. A fixture HTTP server stands in for frank-tts (it answers the
-# speech route with a fixed WAV, the roster route with a JSON array, and
-# everything else with 404) and a stub player stands in for pw-play: it records
-# the path it was given and copies the bytes it was told to play.
+# speech route, (b) the file it hands to the player, and (c) the notifications
+# and select-lists it renders. A fixture HTTP server stands in for frank-tts
+# (it answers the speech route with a fixed WAV, the roster route with a JSON
+# array, and everything else with 404), a stub player stands in for pw-play
+# (it records its full argv and copies the .wav it was given), and a fixture
+# `wpctl` in $WORK/bin stands in for wireplumber's sink enumeration.
 #
 # Driven by env: DV_EXT (extension under test), DV_STUB / DV_STUB_LOG / DV_COPY
-# (player fixture), DV_MODELS (models.yml fixture), DV_EMPTY_BIN (a PATH with no
-# players), DV_MODE (the harness mode reported to the extension).
+# (player fixture), DV_PWPLAY (stub aliased as pw-play), DV_WPCTL (the sink
+# name the fixture wpctl reports), DV_MODELS (models.yml fixture),
+# DANKVOICE_CONFIG (settings file), DV_PLAYER_BIN (player command override),
+# DV_SELECT (pipe-separated answers fed to ctx.ui.select in order), DV_MODE
+# (the harness mode reported to the extension).
 
 setup() {
   REPO="$(cd "$BATS_TEST_DIRNAME/.." && pwd)"
@@ -26,12 +30,27 @@ setup() {
 
   cat > "$DV_STUB" <<'SH'
 #!/usr/bin/env bash
-printf '%s\n' "$1" >> "$DV_STUB_LOG"
-cp "$1" "$DV_COPY"
+printf '%s\n' "$*" >> "$DV_STUB_LOG"
+for arg in "$@"; do case "$arg" in *.wav) cp "$arg" "$DV_COPY" ;; esac; done
 sleep 0.1
 SH
   chmod +x "$DV_STUB"
 
+  # the same stub reachable under the canonical player names, plus a fixture
+  # wpctl so output-device enumeration is deterministic on any machine
+  mkdir -p "$WORK/bin"
+  ln -s "$DV_STUB" "$WORK/bin/pw-play"
+  export DV_PWPLAY="$WORK/bin/pw-play"
+  export DV_WPCTL="alsa_output.usb-KT.0"
+  cat > "$WORK/bin/wpctl" <<'WPC'
+#!/usr/bin/env bash
+printf '57\talsa_output.pci-0000_07_00.6.HiFi__Speaker__sink\taudio/sink\t\n116\t%s\taudio/sink\t*\n' "$DV_WPCTL"
+WPC
+  chmod +x "$WORK/bin/wpctl"
+
+  # settings live in an isolated file so menu choices never leak between tests
+  export DANKVOICE_CONFIG="$WORK/settings.json"
+
   # models.yml fixture: a chat provider first, the speech provider second
   export DV_MODELS="$WORK/models.yml"
   cat > "$DV_MODELS" <<'YML'
@@ -49,11 +68,19 @@ providers:
         api: openai-speech
 YML
 
+  # omp config.yml fixture: the @tiny role points at the chat provider above
+  export DV_OMPCONFIG="$WORK/ompconfig.yml"
+  cat > "$DV_OMPCONFIG" <<'CFY'
+modelRoles:
+  tiny: chat (box)/chatmodel:off
+CFY
+
   cat > "$WORK/driver.ts" <<'TS'
 // Black-box driver: stand up the frank-tts fixture, load the extension against
 // it, invoke one scenario, print what the endpoint received and what got played.
 import { createServer } from "node:http";
 import { existsSync, readFileSync, statSync } from "node:fs";
+import { dirname } from "node:path";
 
 const scenario = process.argv[2] ?? "speak";
 const AUDIO = Buffer.from([0x52, 0x49, 0x46, 0x46, 8, 0, 0, 0, 0x57, 0x41, 0x56, 0x45, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]);
@@ -75,6 +102,14 @@ const server = createServer((req, res) => {
       res.end(AUDIO);
       return;
     }
+    if (process.env.DV_REWRITE && req.url?.endsWith("/chat/completions")) {
+      res.setHeader("content-type", "application/json");
+      const spoken = process.env.DV_REWRITE === "echo"
+        ? String(JSON.parse(body).messages?.[1]?.content ?? "")
+        : "Spoken, humanized words.";
+      res.end(JSON.stringify({ choices: [{ message: { content: spoken } }] }));
+      return;
+    }
     res.statusCode = 404;
     res.end("{}");
   });
@@ -87,8 +122,11 @@ if (scenario === "endpoint") {
 } else {
   process.env.DANKVOICE_URL = `http://127.0.0.1:${port}/v1`;
 }
+process.env.DANKVOICE_OMPCONFIG = process.env.DV_OMPCONFIG;
+if (scenario !== "endpoint") process.env.DANKVOICE_REWRITE_URL = `http://127.0.0.1:${port}/v1`;
 if (scenario === "noplayer") process.env.PATH = process.env.DV_EMPTY_BIN ?? "/nonexistent-bin";
-process.env.DANKVOICE_PLAYER = scenario === "noplayer" ? "" : process.env.DV_STUB;
+else process.env.PATH = `${dirname(process.env.DV_STUB ?? ".")}/bin:${process.env.PATH ?? ""}`;
+process.env.DANKVOICE_PLAYER = scenario === "noplayer" ? "" : process.env.DV_PLAYER_BIN ?? process.env.DV_STUB;
 
 const extension = await import(process.env.DV_EXT as string);
 const commands = new Map<string, { handler: (args: string, ctx: unknown) => Promise<void> }>();
@@ -97,11 +135,19 @@ extension.default({
   registerCommand: (name: string, options: { handler: (args: string, ctx: unknown) => Promise<void> }) => commands.set(name, options),
   on: (event: string, handler: (event: unknown, ctx: unknown) => Promise<void> | void) => events.set(event, handler),
 });
+const answers = (process.env.DV_SELECT ?? "").split("|").filter((answer) => answer !== "");
+let answerIndex = 0;
 const ctx = {
   mode: process.env.DV_MODE ?? "tui",
   ui: {
     notify: (message: string, level?: string) => console.log(`NOTIFY ${level ?? "info"} ${message}`),
     setStatus: (key: string, text?: string) => console.log(`STATUS ${key} ${text ?? ""}`),
+    select: (title: string, options: string[]) => {
+      console.log(`SELECT ${title} ${options.join("|")}`);
+      const answer = answerIndex < answers.length ? answers[answerIndex] : undefined;
+      answerIndex += 1;
+      return Promise.resolve(answer ?? options[0]);
+    },
   },
 };
 const say = commands.get("say");
@@ -121,6 +167,16 @@ if (!say) {
       await say.handler("voice narrator2", ctx);
       await say.handler("good morning", ctx);
       break;
+    case "menu":
+      await say.handler("menu", ctx);
+      break;
+    case "empty":
+      await say.handler("", ctx);
+      break;
+    case "devicemenu":
+      await say.handler("menu", ctx);
+      await say.handler("hello world", ctx);
+      break;
     case "auto":
       await say.handler("on", ctx);
       await agentEnd?.(answer("the answer"), ctx);
@@ -150,6 +206,23 @@ if (!say) {
     case "truncate":
       await say.handler("Sentence one is here. ".repeat(40), ctx);
       break;
+    case "sentences":
+      await say.handler("First stop here. Second one goes!!! Really? Third ... done", ctx);
+      break;
+    case "decimals":
+      await say.handler("Model 3.5 ships today. Version 1,5 too.", ctx);
+      break;
+    case "rewrite":
+      await say.handler("The deploy failed because a secret was missing.", ctx);
+      break;
+    case "thinking":
+      await say.handler("on", ctx);
+      await agentEnd?.({ messages: [{ role: "assistant", content: [{ type: "thinking", thinking: "Considering the wiring." }, { type: "text", text: "The answer is nine." }] }] }, ctx);
+      break;
+    case "narration":
+      await say.handler("on", ctx);
+      await agentEnd?.({ messages: [{ role: "assistant", content: [{ type: "thinking", thinking: "Here is the result.", narration: true }] }] }, ctx);
+      break;
     case "noplayer":
       await say.handler("speak anyway", ctx);
       break;
@@ -168,9 +241,10 @@ for (const call of received) {
   };
   console.log(`CALL ${call.url} model=${field("model")} format=${field("response_format")} voice=${field("voice")} input=${field("input")}`);
 }
-console.log(`CALLS ${received.length}`);
+console.log(`CALLS ${received.filter((call) => call.url?.endsWith("/audio/speech")).length}`);
 const log = process.env.DV_STUB_LOG && existsSync(process.env.DV_STUB_LOG) ? readFileSync(process.env.DV_STUB_LOG, "utf8").trim() : "";
 console.log(`PLAYS ${log === "" ? 0 : log.split("\n").length}`);
+if (log !== "") for (const line of log.split("\n")) console.log(`PLAYER ${line}`);
 if (process.env.DV_COPY && existsSync(process.env.DV_COPY)) console.log(`COPIED ${statSync(process.env.DV_COPY).size}`);
 server.close();
 TS
@@ -183,7 +257,7 @@ teardown() {
 }
 
 say_run() {
-  run bun "$DV_DRIVER" "$1"
+  run env DV_SELECT="${2:-}" bun "$DV_DRIVER" "$1"
 }
 
 @test "the extension registers the say command and an agent_end listener" {
@@ -235,18 +309,76 @@ say_run() {
 @test "the endpoint is taken from the speech row of the models file, never the chat provider" {
   say_run endpoint
   [[ "$output" == *"http://127.0.0.1:1/v1/audio/speech"* ]]
+  [[ "$output" == *"humanize chatmodel"* ]]
   [[ "$output" != *"10.0.0.1"* ]]
 }
 
 @test "markdown never reaches the speech engine" {
   say_run markdown
-  [[ "$output" == *"input=Title Use pw-play here. see docs and bold a b"* ]]
+  [[ "$output" == *"input=Title"* ]]
+  [[ "$output" == *"input=Use pw-play here."* ]]
+  [[ "$output" == *"input=see docs and bold"* ]]
+  [[ "$output" != *"npm i x"* ]]
+  [[ "$output" != *"example.test"* ]]
 }
 
-@test "an over-long utterance is bounded and cut at a sentence boundary" {
+@test "the spoken budget keeps only the leading whole segments" {
   DANKVOICE_MAX=60 say_run truncate
-  [[ "$output" == *"here. …"* ]]
-  [ "$(grep -o 'Sentence one is here\.' <<<"$output" | wc -l)" -le 4 ]
+  [[ "$output" == *"CALLS 1"* ]]
+  [[ "$output" == *"PLAYS 1"* ]]
+  [[ "$output" == *"input=Sentence one is here."* ]]
+}
+
+@test "a long answer is spoken as sequential sentence-sized requests" {
+  say_run sentences
+  [[ "$output" == *"CALLS 3"* ]]
+  [[ "$output" == *"PLAYS 3"* ]]
+  [[ "$output" == *"input=First stop here."* ]]
+  [[ "$output" == *"input=Second one goes!!! Really?"* ]]
+  [[ "$output" == *"input=Third ... done"* ]]
+}
+
+@test "decimals and comma-decimals never split a sentence" {
+  say_run decimals
+  [[ "$output" == *"CALLS 2"* ]]
+  [[ "$output" == *"input=Model 3.5 ships today."* ]]
+  [[ "$output" == *"input=Version 1,5 too."* ]]
+}
+
+@test "an answer is rewritten into spoken prose before synthesis" {
+  DV_REWRITE=prose say_run rewrite
+  [[ "$output" == *"CALL /v1/chat/completions"* ]]
+  [[ "$output" == *"input=Spoken, humanized words."* ]]
+  [[ "$output" == *"CALLS 1"* ]]
+}
+
+@test "a failed rewrite speaks the mechanical text unchanged" {
+  say_run speak
+  [[ "$output" == *"input=hello world"* ]]
+}
+
+@test "thinking is narrated before the answer as its own segment" {
+  DV_REWRITE=echo say_run thinking
+  [[ "$output" == *"input=Considering the wiring."* ]]
+  [[ "$output" == *"input=The answer is nine."* ]]
+  [[ "$output" == *"CALLS 2"* ]]
+  local first second
+  first=$(grep -n "input=Considering the wiring\." <<<"$output" | head -1 | cut -d: -f1)
+  second=$(grep -n "input=The answer is nine\." <<<"$output" | head -1 | cut -d: -f1)
+  [ -n "$first" ] && [ -n "$second" ] && [ "$first" -lt "$second" ]
+}
+
+@test "thinking narration can be switched off" {
+  DANKVOICE_THINKING=0 DV_REWRITE=echo say_run thinking
+  [[ "$output" == *"input=The answer is nine."* ]]
+  [[ "$output" != *"input=Considering the wiring."* ]]
+  [[ "$output" == *"CALLS 1"* ]]
+}
+
+@test "narration-flagged thinking is spoken once as the answer" {
+  DV_REWRITE=echo say_run narration
+  [[ "$output" == *"input=Here is the result."* ]]
+  [[ "$output" == *"CALLS 1"* ]]
 }
 
 @test "a missing player is reported, not swallowed" {
@@ -261,3 +393,27 @@ say_run() {
  [[ "$output" == *"input=the settled answer"* ]]
  [[ "$output" != *"input=not yet"* ]]
 }
+
+@test "the menu lists every knob and a plain /say opens it" {
+  say_run menu
+  [[ "$output" == *"SELECT dankvoice — options Auto-speak: off (toggle)|Voice: server default (change)|Output device: system default (change)|Player: $DV_STUB (change)|Spoken cap: 700 chars (change)|Thinking narration: on (toggle)|Humanize answers: on (tiny) (toggle)|Speak a test phrase"* ]]
+  [[ "$output" == *"NOTIFY info dankvoice: auto-speak armed (server default @ http://127.0.0.1:"* ]]
+  say_run empty
+  [[ "$output" == *"SELECT dankvoice — options"* ]]
+}
+
+@test "a voice picked from the menu is spoken by every later request" {
+  say_run menu "Voice: server default (change)|narrator2"
+  [[ "$output" == *"SELECT dankvoice — voice server default|narrator1|narrator2"* ]]
+  [[ "$output" == *"NOTIFY info dankvoice: voice = narrator2"* ]]
+  say_run speak
+  [[ "$output" == *"voice=narrator2 input=hello world"* ]]
+}
+
+@test "a device chosen from the menu becomes the player's output target" {
+  DV_PLAYER_BIN="$DV_PWPLAY" say_run menu "Output device: system default (change)|$DV_WPCTL"
+  [[ "$output" == *"SELECT dankvoice — output device system default|alsa_output.pci-0000_07_00.6.HiFi__Speaker__sink|$DV_WPCTL"* ]]
+  [[ "$output" == *"NOTIFY info dankvoice: output device = $DV_WPCTL"* ]]
+  DV_PLAYER_BIN="$DV_PWPLAY" say_run speak
+  [[ "$output" == *"PLAYER --target $DV_WPCTL"* ]]
+}