@hicaru/pi-rlm 0.1.0 → 0.1.2

package/README.md

@@ -1,25 +1,40 @@
+# pi-rlm — Save 99% tokens, Recursive Language Model (RLM) for the Pi
+
 <div align="center">
 
-<img src="../../assets/hero.png" alt="pi-rlm">
+**Recursive Language Models (RLMs)**, implemented natively as a Pi extension —
+FULLY LOCAL.
 
 </div>
 
-<div align="center">
+## Install
 
-<sub>
-**English** &nbsp;·&nbsp; <a href="README.zh-CN.md">中文</a> &nbsp;·&nbsp; <a href="README.ru.md">Русский</a>
-</sub>
+```bash
+pi install npm:@hicaru/pi-rlm
+```
 
-</div>
+To remove it later:
 
----
+```bash
+pi uninstall npm:@hicaru/pi-rlm
+```
+
+Then run `/reload` or restart Pi. Verify with `pi list` that the package appears in
+`settings.packages`, and check that `/rlm`, `/rlm-config`, and `/rlm-stop` appear under **[Extensions]**.
+
+<div align="center">
+
+<a href="https://arxiv.org/abs/2512.24601"><img src="../../assets/hero.png" alt="pi-rlm"></a>
 
-# pi-rlm — Recursive Language Models for the [Pi](https://github.com/earendil-works) Coding Agent
+<sub>Modeled on the method in the RLM paper, reimplemented natively for Pi.</sub>
+
+</div>
 
 <div align="center">
 
-**Recursive Language Models (RLMs)**, implemented natively as a Pi extension —
-no extra servers, no Docker, no sockets.
+<sub>
+**English** &nbsp;·&nbsp; <a href="README.zh-CN.md">中文</a> &nbsp;·&nbsp; <a href="README.ru.md">Русский</a>
+</sub>
 
 </div>
 
@@ -46,59 +61,31 @@ sub-LLM calls, hence the name.
 - Hard sub-problems **recurse** into child RLMs via `rlm_query` (depth-capped).
 - Everything runs **in-process** — the only external process is one local `python3` worker.
 
-> This is a Pi-plugin reimplementation of the RLM method (see the [RLM paper](https://arxiv.org/abs/2512.24601)
-> and the [Python `rlm` library](https://github.com/alexzhang13/rlm-minimal)). It is **not** the Python library.
+> This is a Pi-plugin reimplementation of the RLM method (see the [RLM paper](https://arxiv.org/abs/2512.24601)).
+> It is **not** the Python library.
 
 ## How it works
 
 ```
-pi process (TypeScript)
- ├─ /rlm  ──► engine drives the SMART (root) model turn-by-turn (writes ```repl``` Python)
- │             │  each turn: parse repl blocks ──► run in sandbox ──► feed stdout back
- │             ▼
- ├─ bridge ── llm_query / llm_query_batched ──► WORKER model (serverless, in-process)
- │            rlm_query ──► recursive child RLM (own sandbox), depth-capped
- ├─ AgentTree ──► live agent/subagent tree above the editor (roles, depth, cost, tokens)
- └─ PythonSandbox ── `python3 worker.py` ──[JSONL over stdio, bidirectional]── persistent REPL
-```
-
-- **No servers, no sockets, no Docker.** The only external process is one local `python3` sandbox.
-  When sandbox code calls `llm_query`, the worker writes a request on stdout and blocks on stdin;
-  Pi services it in-process and writes the reply back. **Provider API keys never enter the sandbox.**
-- The sandbox exposes `context`, `llm_query`, `llm_query_batched`, `rlm_query`,
-  `rlm_query_batched`, `SHOW_VARS()`, `todo()`, `ask_user_question()`, and an `answer` dict.
-  The model submits its final result by setting `answer["ready"] = True`.
-
-## Install
-
-`pi-rlm` is a Pi package. Pi provides the `@earendil-works/pi-*` and `typebox` peer
-dependencies; do **not** install a separate copy of them into this package. Requires
-`python3` on `PATH` (standard library only).
-
-Recommended local install while developing:
-
-```bash
-pi install /path/to/this-repo/pi-plugin/rlm
-```
-
-Published npm package install:
-
-```bash
-npm publish                       # e.g. as @<you>/pi-rlm
-pi install npm:@<you>/pi-rlm
+          ┌─────────────────────────┐
+          │     Pi coding agent     │
+          └────────────┬────────────┘
+                       │  /rlm
+                       ▼
+          ┌─────────────────────────┐  spawns   ┌────────────────────┐
+          │  Smart model (root)     │ ────────►  │   Worker models    │
+          │  drives a Python REPL   │ ◄────────  │   (cheap, fast)    │
+          └────────────┬────────────┘  results  └────────────────────┘
+                       │ recursion (depth-capped)
+                       └────► child RLMs ────► (same loop)
+
+   All local · one python3 process · no servers
 ```
 
-> **Git installs** require the package manifest to live at the installed repository root.
-> For monorepo subdirectories like this one, prefer the local-path or npm flow above.
-
-If you previously copied the extension folder directly, remove it so it does not shadow the package:
-
-```bash
-rm -rf ~/.pi/agent/extensions/rlm
-```
-
-Then run `/reload` or restart Pi. Verify with `pi list` that the package appears in
-`settings.packages`, and check that `/rlm`, `/rlm-config`, and `/rlm-stop` appear under **[Extensions]**.
+- The **smart model** thinks and writes Python in a REPL.
+- The **worker models** do the heavy lifting (read, summarize, classify).
+- Hard sub-problems **recurse** into child RLMs.
+- Everything runs **fully local** — your API keys never leave Pi.
 
 ## Commands
 
@@ -129,6 +116,8 @@ These functions are injected into the model's Python namespace inside the REPL:
 | `rlm_query_batched` | `(prompts, model=None) -> list[str]` | Concurrent recursive child RLMs |
 | `todo` | `(action, **kwargs) -> str` | Task list: `create`/`update`/`list`/`get`/`delete`/`clear` |
 | `ask_user_question` | `(questions) -> list[dict]` | Ask the user structured questions (depth 0 only) |
+| `stage_edit` | `(path, old_text, new_text) -> str` | Stage a file edit; relayed to the host's native edit flow |
+| `advance_phase` | `(phase, summary=None) -> str` | Move the root pipeline to a new phase |
 | `SHOW_VARS` | `() -> str` | List currently defined variables & their types |
 | `answer` | `dict` | Set `answer["content"]=...; answer["ready"]=True` to finalize |
 
@@ -138,15 +127,18 @@ These functions are injected into the model's Python namespace inside the REPL:
 |---|---|---|
 | Smart model | Pi's active model | the root orchestrator |
 | Worker model | cheapest available | answers `llm_query` |
-| Max recursion depth | `4` | `rlm_query` past this falls back to `llm_query` |
-| Max iterations | `30` | turns before the engine finalizes |
-| Budget ceiling | none | stops the whole tree when USD spend exceeds this |
-| Max consecutive errors | `5` | stops after N consecutive error turns |
-| REPL block timeout | `120s` | per-`repl`-block wall-clock (SIGALRM in the worker) |
-| Max concurrent sub-calls | `4` | pool size for `*_batched` |
-| Orchestrator addendum | on | "delegate, don't solve" guidance |
-| Trajectory compaction | on (0.85) | summarize history when it nears the context window |
-| `yolo` | off | apply proposed edits immediately, skipping the review popup |
+| Max recursion depth | `4` | `rlm_query` past this degrades to plain `llm_query` |
+| Max iterations | `30` | root REPL turns before RLM asks for a final answer |
+| REPL block timeout (s) | `120` | wall-clock limit for one Python REPL block (SIGALRM) |
+| Max concurrent sub-calls | `4` | concurrency pool size for `*_batched` |
+| Budget ceiling (USD) | none | total spend cap for the whole recursive tree |
+| Wall-clock ceiling (min) | none | total runtime cap for the whole recursive tree |
+| Token ceiling | none | total input+output token cap for the whole recursive tree |
+| Max consecutive errors | `5` | stop after N consecutive failing turns (none = off) |
+| Orchestrator addendum | on | divide-and-conquer guidance in the root system prompt |
+| Trajectory compaction | on (0.65) | summarize old turns when history nears the context window |
+| Root model output cap (tok) | `16384` | max output tokens per root-model turn |
+| Sandbox init timeout | `30000` ms | how long to wait for the Python worker to start |
 | `askUserQuestion` | on | expose `ask_user_question()` to the model |
 | `todo` | on | expose `todo()` to the model |
 
@@ -155,17 +147,6 @@ These functions are injected into the model's Python namespace inside the REPL:
 > defaults (depth 4, conc 4) that's 4³ = 64 in the pathological case. Budget and error
 > caps (above) bound total spend regardless of fan-out.
 
-## Telemetry & run logs
-
-- **Run logs** (`runLog`): always-on by default. Each run writes a JSONL trail to `.rlm/runs/`
-  (default), capped at `maxRuns` (50). Supports **snapshots** (`sandbox.pkl`) and **resume**
-  of interrupted runs via `/rlm-resume`. Snapshots are protected by a per-session `nonce`
-  to prevent cross-session replay.
-- **MLflow tracing** (`telemetry`): optional. Set `MLFLOW_TRACKING_URI` or configure
-  `trackingUri` / `experimentId` in `/rlm-config`. The root run is tagged as an MLflow span
-  for trace correlation on resume. The Bearer token comes from the `MLFLOW_TRACKING_TOKEN`
-  env var and is **never persisted** to `rlm.json`.
-
 ## Security
 
 - **Key isolation**: provider keys live only in TypeScript (`AuthStorage`); the sandbox
@@ -181,57 +162,3 @@ These functions are injected into the model's Python namespace inside the REPL:
   SIGALRM timeout + parent watchdog (SIGKILL on hang); budget / token / timeout /
   consecutive-error caps.
 - **Trust**: project-local install requires Pi project trust.
-
-## Project layout
-
-```
-src/
-  sandbox/    worker.py + JSONL stdio driver (PythonSandbox) · protocol.ts · sandbox-manager.ts
-  bridge/     model.ts (one-shot completion) · llm-query.ts · rlm-query.ts (recursion)
-  core/       engine.ts (the loop) · iteration · limits · answer · compaction · pipeline · types
-  prompts/    system + per-turn prompts (ported from the Python reference)
-  text/       parsing (repl blocks) · tokens · preview · edits
-  state/      agent-tree · events · reads/writes · resume · paths · rows
-  tool/       repl-tool · rlm-events · aggregator · propose-edits · emitter-listener
-  config/     defaults · settings (rlm.json persistence + validation)
-  context/    repomix-based repository packing + caching
-  telemetry/  MLflow sink · dispatcher · mlflow-config
-  ui/         tree-widget · status · model-picker · config-panel · intro · theme
-  commands/   rlm · rlm-config
-  mode/       rlm-mode (controller) · input-router
-  patch/      apply · popup · index
-  util/       errors · concurrency
-test/         phase1–phase9 · native-smoke · native-mode · helpers
-```
-
-## Tests
-
-Runtime is **Bun** (`bun install`, `bun run …` — never npm/pnpm/yarn).
-
-```bash
-bun run test/phase1.ts                   # sandbox: exec, persistence, key isolation, timeout kill
-bun run test/phase4.ts                   # recursion depth-cap logic (no tokens)
-bun run test/phase5.ts                   # live agent tree rendering (no tokens)
-RLM_TEST_LIVE=1 bun run test/phase2.ts   # real llm_query through the sandbox
-RLM_TEST_LIVE=1 bun run test/phase3.ts   # real end-to-end /rlm over a file context
-RLM_TEST_LIVE=1 bun run test/phase4.ts   # engine solves a 20-doc needle-in-haystack
-```
-
-## Background
-
-Modeled on the Python reference [`rlm`](https://github.com/alexzhang13/rlm-minimal) and the
-method in the [RLM paper](https://arxiv.org/abs/2512.24601), reimplemented natively for Pi.
-
-If you use this in your research, please cite the original RLM work:
-
-```bibtex
-@misc{zhang2026recursivelanguagemodels,
-      title={Recursive Language Models},
-      author={Alex L. Zhang and Tim Kraska and Omar Khattab},
-      year={2026},
-      eprint={2512.24601},
-      archivePrefix={arXiv},
-      primaryClass={cs.AI},
-      url={https://arxiv.org/abs/2512.24601},
-}
-```

package/README.ru.md

@@ -19,7 +19,7 @@
 <div align="center">
 
 **Рекурсивные языковые модели (RLMs)**, реализованные нативно как расширение Pi —
-без дополнительных серверов, Docker или сокетов.
+ПОЛНОСТЬЮ ЛОКАЛЬНО.
 
 </div>
 
@@ -36,8 +36,8 @@
 - Сложные подзадачи **рекурсивно** передаются в дочерние RLM через `rlm_query` (с ограничением глубины).
 - Все работает **in-process** — единственным внешним процессом является локальный worker `python3`.
 
-> This is a Pi-plugin reimplementation of the RLM method (see the [RLM paper](https://arxiv.org/abs/2512.24601)
-> and the [Python `rlm` library](https://github.com/alexzhang13/rlm-minimal)). It is **not** the Python library.
+> This is a Pi-plugin reimplementation of the RLM method (see the [RLM paper](https://arxiv.org/abs/2512.24601)).
+> It is **not** the Python library.
 
 ## Как это работает
 
@@ -133,10 +133,9 @@ rm -rf ~/.pi/agent/extensions/rlm
 
 > **Примечание по параллелизму:** каждый дочерний `rlm_query` запускает собственного worker `python3` (~50–150 мс «холодного старта»). В худшем случае количество параллельных интерпретаторов ≈ `maxConcurrentSubcalls`^(depth−1); при настройках по умолчанию (глубина 4, параллелизм 4) это 4³ = 64 в патологическом случае. Лимиты бюджета и ошибок (см. выше) ограничивают общие затраты независимо от степени разветвления.
 
-## Телеметрия и логи запусков
+## Логи запусков
 
 - **Логи запусков** (`runLog`): включены по умолчанию. Каждый запуск записывает след в формате JSONL в `.rlm/runs/` (по умолчанию) с ограничением `maxRuns` (50). Поддерживает **снимки** (`sandbox.pkl`) и **возобновление** прерванных запусков через `/rlm-resume`. Снимки защищены сессионным `nonce` для предотвращения повторов между сессиями.
-- **Трассировка MLflow** (`telemetry`): опционально. Установите `MLFLOW_TRACKING_URI` или настройте `trackingUri` / `experimentId` в `/rlm-config`. Корневой запуск помечается как span MLflow для корреляции трасс при возобновлении. Bearer-токен берется из переменной окружения `MLFLOW_TRACKING_TOKEN` и **никогда не сохраняется** в `rlm.json`.
 
 ## Безопасность
 
@@ -155,11 +154,10 @@ src/
   core/       engine.ts (цикл) · iteration · limits · answer · compaction · pipeline · types
   prompts/    системные промпты и промпты для каждого шага (перенесены из Python-референса)
   text/       парсинг (repl-блоки) · токены · превью · правки
-  state/      дерево-агентов · события · чтения/записи · возобновление · пути · строки
+  state/      чтения/записи · возобновление · пути · строки
   tool/       repl-tool · rlm-events · агрегатор · предложение-правок · emitter-listener
   config/     значения по умолчанию · настройки (сохранение и валидация rlm.json)
   context/    упаковка репозитория на базе repomix + кеширование
-  telemetry/  MLflow sink · диспетчер · mlflow-config
   ui/         виджет-дерева · статус · выбор-модели · панель-конфигурации · вступление · тема
   commands/   rlm · rlm-config
   mode/       rlm-mode (контроллер) · маршрутизатор-ввода
@@ -183,7 +181,7 @@ RLM_TEST_LIVE=1 bun run test/phase4.ts   # engine solves a 20-doc needle-in-hays
 
 ## Общая информация
 
-Реализовано на основе эталонного проекта [`rlm`](https://github.com/alexzhang13/rlm-minimal) на Python и метода из [статьи RLM](https://arxiv.org/abs/2512.24601), с нативной переработкой для Pi.
+Реализовано на основе метода из [статьи RLM](https://arxiv.org/abs/2512.24601), с нативной переработкой для Pi.
 
 Если вы используете этот проект в своих исследованиях, пожалуйста, сошлитесь на оригинальную работу RLM:
 

package/README.zh-CN.md

@@ -19,7 +19,7 @@
 <div align="center">
 
 **递归语言模型 (RLMs)** 作为 Pi 扩展原生实现 ——
-无需额外服务器，无需 Docker，无需 socket。
+完全本地。
 
 </div>
 
@@ -36,8 +36,8 @@
 - 困难的子问题通过 `rlm_query` **递归**到子 RLM 中（设有深度限制）。
 - 所有内容均**在进程内**运行 —— 唯一的外部进程是一个本地的 `python3` worker。
 
-> 这是 RLM 方法的 Pi 插件重新实现（参见 [RLM 论文](https://arxiv.org/abs/2512.24601)
-> 和 [Python `rlm` 库](https://github.com/alexzhang13/rlm-minimal)）。它**不是**那个 Python 库。
+> 这是 RLM 方法的 Pi 插件重新实现（参见 [RLM 论文](https://arxiv.org/abs/2512.24601)）。
+> 它**不是**那个 Python 库。
 
 ## 工作原理
 
@@ -143,15 +143,11 @@ rm -rf ~/.pi/agent/extensions/rlm
 > 默认设置下 (深度 4, 并发 4)，极端情况下为 4³ = 64。预算和错误
 > 上限 (见上文) 无论扇出 (fan-out) 如何都会限制总支出。
 
-## 遥测与运行日志
+## 运行日志
 
 - **运行日志** (`runLog`)：默认始终开启。每次运行将 JSONL 轨迹写入 `.rlm/runs/`
   (默认)，上限为 `maxRuns` (50)。支持通过 `/rlm-resume` 进行**快照** (`sandbox.pkl`) 和**恢复**
   被中断的任务。快照受每个会话的 `nonce` 保护，以防止跨会话重放。
-- **MLflow 追踪** (`telemetry`)：可选。设置 `MLFLOW_TRACKING_URI` 或在
-  `/rlm-config` 中配置 `trackingUri` / `experimentId`。根运行被标记为 MLflow span
-  以便在恢复时进行追踪关联。Bearer 令牌来自 `MLFLOW_TRACKING_TOKEN`
-  环境变量，且**绝不会**持久化到 `rlm.json`。
 
 ## 安全性
 
@@ -178,11 +174,10 @@ src/
   core/       engine.ts (the loop) · iteration · limits · answer · compaction · pipeline · types
   prompts/    system + per-turn prompts (ported from the Python reference)
   text/       parsing (repl blocks) · tokens · preview · edits
-  state/      agent-tree · events · reads/writes · resume · paths · rows
+  state/      reads/writes · resume · paths · rows
   tool/       repl-tool · rlm-events · aggregator · propose-edits · emitter-listener
   config/     defaults · settings (rlm.json persistence + validation)
   context/    repomix-based repository packing + caching
-  telemetry/  MLflow sink · dispatcher · mlflow-config
   ui/         tree-widget · status · model-picker · config-panel · intro · theme
   commands/   rlm · rlm-config
   mode/       rlm-mode (controller) · input-router
@@ -206,8 +201,7 @@ RLM_TEST_LIVE=1 bun run test/phase4.ts   # 引擎解决 20 个文档的“大海
 
 ## 背景
 
-基于 Python 参考实现 [`rlm`](https://github.com/alexzhang13/rlm-minimal) 和
-[RLM 论文](https://arxiv.org/abs/2512.24601) 中的方法，为 Pi 原生重新实现。
+基于 [RLM 论文](https://arxiv.org/abs/2512.24601) 中的方法，为 Pi 原生重新实现。
 
 如果您在研究中使用此项目，请引用原始 RLM 工作：
 

package/package.json

@@ -1,8 +1,8 @@
 {
   "name": "@hicaru/pi-rlm",
-  "version": "0.1.0",
+  "version": "0.1.2",
   "type": "module",
-  "description": "Recursive Language Model (RLM) for the Pi coding agent — native, server-less.",
+  "description": "Save 99% tokens, Recursive Language Model (RLM) for the Pi",
   "license": "MIT",
   "author": "hicaru",
   "repository": {
@@ -13,17 +13,17 @@
   "bugs": {
     "url": "https://github.com/hicaru/rlm.pi/issues"
   },
-  "exports": {
-    ".": "./src/index.ts"
-  },
   "files": [
-    "src",
+    "src/",
     "README.md",
     "LICENSE"
   ],
   "keywords": [
     "pi-package",
-    "pi-extension"
+    "pi-extension",
+    "rlm",
+    "recursive",
+    "ai-agent"
   ],
   "scripts": {
     "check": "tsc --noEmit",
@@ -38,14 +38,12 @@
     "access": "public"
   },
   "peerDependencies": {
-    "@earendil-works/pi-ai": ">=0.79.0",
-    "@earendil-works/pi-coding-agent": ">=0.79.0",
-    "@earendil-works/pi-tui": ">=0.79.0",
+    "@earendil-works/pi-ai": "*",
+    "@earendil-works/pi-coding-agent": "*",
+    "@earendil-works/pi-tui": "*",
     "typebox": "*"
   },
   "dependencies": {
-    "@mlflow/core": "~0.2.0",
-    "diff": "^9.0.0",
     "repomix": "^1.15.0"
   },
   "devDependencies": {

package/src/bridge/pi-interactive.ts

@@ -4,29 +4,6 @@ import type { AskAnswer, AskQuestion } from "../sandbox/protocol.ts";
 import { formatError } from "../util/errors.ts";
 import { createTodoFallback } from "./fallback-todo.ts";
 
-interface ToolInvoker {
-  readonly callTool?: (name: string, params: unknown) => Promise<unknown>;
-}
-
-function hasAnswers(value: unknown): value is { readonly answers: readonly unknown[] } {
-  return typeof value === "object" && value !== null && Array.isArray((value as { readonly answers?: unknown }).answers);
-}
-
-function isAskAnswer(value: unknown): value is AskAnswer {
-  if (typeof value !== "object" || value === null) return false;
-  const candidate = value as { readonly question?: unknown; readonly selected?: unknown; readonly custom?: unknown };
-  return typeof candidate.question === "string"
-    && Array.isArray(candidate.selected)
-    && candidate.selected.every((item) => typeof item === "string")
-    && (candidate.custom === undefined || typeof candidate.custom === "string");
-}
-
-function normalizeAnswers(result: unknown): AskAnswer[] | undefined {
-  if (!hasAnswers(result)) return undefined;
-  const answers = result.answers;
-  return answers.every(isAskAnswer) ? Array.from(answers) : undefined;
-}
-
 async function askViaUi(ctx: ExtensionContext, questions: readonly AskQuestion[]): Promise<AskAnswer[]> {
   if (!ctx.hasUI) throw new Error("ask_user_question requires UI");
   const answers = new Array<AskAnswer>(questions.length);
@@ -57,30 +34,8 @@ async function askViaUi(ctx: ExtensionContext, questions: readonly AskQuestion[]
 export function createPiInteractiveDeps(ctx: ExtensionContext): InteractiveDeps {
   const fallbackTodo = createTodoFallback();
   return Object.freeze({
-    onAskUserQuestion: async (questions: readonly AskQuestion[]): Promise<AskAnswer[]> => {
-      const callTool = (ctx as unknown as ToolInvoker).callTool;
-      if (typeof callTool === "function") {
-        try {
-          const result = await callTool.call(ctx, "ask_user_question", { questions });
-          const answers = normalizeAnswers(result);
-          if (answers) return answers;
-        } catch {
-          // Fall through to native UI fallback when the extension tool is not registered or fails.
-        }
-      }
-      return askViaUi(ctx, questions);
-    },
-    onTodo: async (action: string, params: Record<string, unknown>): Promise<string> => {
-      const callTool = (ctx as unknown as ToolInvoker).callTool;
-      if (typeof callTool === "function") {
-        try {
-          const result = await callTool.call(ctx, "todo", { action, ...params });
-          return typeof result === "string" ? result : JSON.stringify(result);
-        } catch {
-          // Fall through to in-process task store when the extension tool is not registered or fails.
-        }
-      }
-      return fallbackTodo(action, params);
-    },
+    onAskUserQuestion: (questions: readonly AskQuestion[]): Promise<AskAnswer[]> => askViaUi(ctx, questions),
+    onTodo: (action: string, params: Record<string, unknown>): Promise<string> =>
+      Promise.resolve(fallbackTodo(action, params)),
   });
 }

package/src/commands/rlm-config.ts

@@ -11,15 +11,22 @@ export async function runRlmConfig(controller: RlmController, ctx: ExtensionCont
   const models = ctx.modelRegistry.getAvailable();
 
   const worker = await selectModel(ctx, "Worker model (sub-LLM / llm_query)", models, controller.workerModel, controller.config.subSampling.reasoning);
-  if (worker) {
+  if (worker === null) {
+    controller.workerModel = undefined;
+    controller.config.subSampling.reasoning = undefined;
+  } else if (worker) {
     controller.workerModel = worker.model;
     controller.config.subSampling.reasoning = worker.thinkingLevel;
   }
 
   await showConfigPanel(ctx, controller.config);
 
-  const effectiveWorker = controller.workerModel ?? cheapestModel(ctx.modelRegistry);
-  controller.savedWorkerRef = modelRef(controller.workerModel) ?? modelRef(effectiveWorker);
+  if (worker === null) {
+    controller.savedWorkerRef = undefined;
+  } else {
+    const effectiveWorker = controller.workerModel ?? cheapestModel(ctx.modelRegistry);
+    controller.savedWorkerRef = modelRef(controller.workerModel) ?? modelRef(effectiveWorker);
+  }
   const persisted = await controller.persist();
   if (!persisted) ctx.ui.notify("RLM: failed to save settings to ~/.pi/agent/rlm.json", "error");
   setRlmModeStatus(ctx.ui, controller);

package/src/commands/rlm.ts

@@ -13,9 +13,6 @@ import type { RunHeader } from "../state/rows.ts";
 import { buildRlmSystemPrompt } from "../prompts/system.ts";
 import { RlmEmitter } from "../tool/rlm-events.ts";
 import { RlmEventAggregator } from "../tool/rlm-aggregator.ts";
-import { createTelemetrySink } from "../telemetry/index.ts";
-import { applyEdits } from "../patch/index.ts";
-import { tryExtractDiff } from "../core/answer.ts";
 
 export function registerRlmCommand(pi: ExtensionAPI, controller: RlmController): void {
   pi.registerCommand("rlm", {
@@ -112,14 +109,10 @@ async function executeRlmRunWithResume(
   context: unknown,
 ): Promise<void> {
   let handle: RunHandle | undefined;
-  let sink: Awaited<ReturnType<typeof createTelemetrySink>>;
   let emitter: RlmEmitter | undefined;
-  let detachSink: (() => void) | undefined;
   let aggregator: RlmEventAggregator | undefined;
   try {
-    sink = await createTelemetrySink(controller.config.telemetry);
     emitter = new RlmEmitter();
-    if (sink) detachSink = emitter.attachSink(sink);
     aggregator = new RlmEventAggregator(emitter, (partial) => {
       const d = partial.details;
       if (!d) return;
@@ -146,20 +139,12 @@ async function executeRlmRunWithResume(
   try {
     const result = await done;
     pi.sendMessage({ customType: "rlm-answer", content: result.answer, display: true });
-    const proposedDiffs = result.diffs?.length ? result.diffs : tryExtractDiff(result.answer);
-    await applyEdits(
-      result.edits ?? [],
-      proposedDiffs,
-      ctx,
-    );
   } catch (e) {
     ctx.ui.notify(`RLM resume failed: ${e instanceof Error ? e.message : String(e)}`, "error");
   } finally {
     clearRlmStatus(ctx.ui);
     ctx.ui.setWidget?.("rlm-status", undefined);
-    detachSink?.();
     aggregator?.dispose();
     emitter?.shutdown();
-    try { await sink?.shutdown(); } catch { /* best-effort */ }
   }
 }

package/src/config/settings.ts

@@ -4,7 +4,7 @@ import { mkdir, readFile, writeFile } from "node:fs/promises";
 import { dirname, join } from "node:path";
 import { getAgentDir, type ModelRegistry } from "@earendil-works/pi-coding-agent";
 import type { Api, Model, ThinkingLevel } from "@earendil-works/pi-ai";
-import type { RlmConfig, RunLogConfig, TelemetryConfig } from "../core/types.ts";
+import type { RlmConfig, RunLogConfig } from "../core/types.ts";
 import { DEFAULT_CONFIG } from "./defaults.ts";
 
 export interface PersistedSettings {
@@ -30,26 +30,6 @@ function validateString(v: unknown): string | undefined {
   return typeof v === "string" && v.trim() ? v : undefined;
 }
 
-function validateTelemetry(raw: unknown): TelemetryConfig | undefined {
-  if (typeof raw !== "object" || raw === null) return undefined;
-  const r = raw as Record<string, unknown>;
-  const out: {
-    enabled?: boolean;
-    trackingUri?: string;
-    experimentId?: string;
-    maxQueueSize?: number;
-  } = {};
-  const enabled = validateBoolean(r.enabled);
-  if (enabled !== undefined) out.enabled = enabled;
-  const trackingUri = validateString(r.trackingUri);
-  if (trackingUri !== undefined) out.trackingUri = trackingUri;
-  const experimentId = validateString(r.experimentId);
-  if (experimentId !== undefined) out.experimentId = experimentId;
-  const maxQueueSize = validateNumber(r.maxQueueSize, 1);
-  if (maxQueueSize !== undefined) out.maxQueueSize = maxQueueSize;
-  return Object.freeze(out);
-}
-
 function validateRunLog(raw: unknown): Partial<RunLogConfig> | undefined {
   if (typeof raw !== "object" || raw === null) return undefined;
   const r = raw as Record<string, unknown>;
@@ -102,8 +82,6 @@ function validateConfig(raw: unknown): Partial<RlmConfig> {
   if (typeof r.smartReasoning === "string") out.smartReasoning = r.smartReasoning as ThinkingLevel;
   const subSystemPrompt = validateString(r.subSystemPrompt);
   if (subSystemPrompt !== undefined) out.subSystemPrompt = subSystemPrompt;
-  const telemetry = validateTelemetry(r.telemetry);
-  if (telemetry) out.telemetry = telemetry;
   const runLog = validateRunLog(r.runLog);
   if (runLog) out.runLog = runLog;
   const sandboxInitTimeoutMs = validateNumber(r.sandboxInitTimeoutMs, 100);
@@ -167,7 +145,6 @@ export function mergeConfig(partial: Partial<RlmConfig>): RlmConfig {
     ...partial,
     subSampling: { ...DEFAULT_CONFIG.subSampling, ...partial.subSampling },
     rootSampling: Object.freeze({ ...DEFAULT_CONFIG.rootSampling, ...partial.rootSampling }),
-    ...(partial.telemetry ? { telemetry: Object.freeze({ ...partial.telemetry }) } : {}),
     ...(partial.runLog ? { runLog: Object.freeze({ ...DEFAULT_CONFIG.runLog, ...partial.runLog }) } : {}),
   };
 }

package/src/core/answer.ts

@@ -1,6 +1,6 @@
 /** Helpers for detecting and formatting the RLM final answer from a turn's REPL results. */
 
-import type { ProposedDiffEdit, ProposedEdit, ReplResult } from "../sandbox/protocol.ts";
+import type { ProposedEdit, ReplResult } from "../sandbox/protocol.ts";
 import { truncateOutput } from "../text/parsing.ts";
 
 /** First non-null final answer across a turn's executed blocks, or null. */
@@ -27,22 +27,6 @@ export function collectEdits(results: readonly ReplResult[]): ProposedEdit[] {
   return [];
 }
 
-/** Last cumulative diff proposal set reported by a turn. */
-export function collectDiffs(results: readonly ReplResult[]): ProposedDiffEdit[] {
-  for (let i = results.length - 1; i >= 0; i--) {
-    const diffs = results[i]?.diffs;
-    if (diffs && diffs.length > 0) return [...diffs];
-  }
-  return [];
-}
-
-/** Parses a ```diff fence from answer text; returns [] if none found. */
-export function tryExtractDiff(answer: string): ProposedDiffEdit[] {
-  const match = /```diff\n([\s\S]*?)```/.exec(answer);
-  if (!match) return [];
-  return [{ diff: match[1].trim() }];
-}
-
 /** True if any block in the turn raised an exception. Plain stderr does not count. */
 export function turnHadError(results: readonly ReplResult[]): boolean {
   return results.some((r) => r.raised);