Release v202611-00007

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

CHANGELOG.md

@@ -3,6 +3,28 @@
 All notable changes to CodeCraft. Versions follow [Semantic Versioning](https://semver.org/);
 until 1.0, minor versions may contain breaking changes.
 
+## [v202611-00007] – 2026-11-09
+
+### Added
+
+- **Telegram bot as a chat interface:** every user can connect their own bot (*Settings →
+  Account → Telegram*: token from @BotFather, then `/start <code>` pairs it with their Telegram
+  account – other chats are ignored). Messages, photos and files go to the current session; the
+  agent's steps show live, the answer comes formatted, images from tools come as photos, and
+  approvals and questions have buttons. Commands like the terminal client: `/sessions`,
+  `/session`, `/new`, `/clear`, `/mode`, `/model`, `/permissions`, `/effort`, `/stop`, `/status`,
+  `/costs`, `/memory`, `/jobs`, `/compact`, `/help`. The server polls Telegram – no public
+  webhook needed.
+
+## [v202611-00006] – 2026-11-09
+
+### Added
+
+- The image contains the Docker CLI (static client). With the host's `/var/run/docker.sock`
+  mounted (and its group), admins' agents and terminals can use the host's Docker;
+  `scripts/move-server.sh --docker-sock` sets this up on a new host.
+- `scripts/move-server.sh --tools-dir DIR`: agent home, users' homes and `/tmp` on a separate disk.
+
 ## [v202611-00005] – 2026-11-09
 
 ### Fixed

Cargo.lock

@@ -3040,6 +3040,7 @@ dependencies = [
  "hyper-util",
  "js-sys",
  "log",
+ "mime_guess",
  "percent-encoding",
  "pin-project-lite",
  "quinn",

Dockerfile

@@ -67,6 +67,14 @@ RUN mkdir -p server/src && echo 'fn main(){}' > server/src/main.rs \
     && mkdir -p /out && cp target/x86_64-unknown-linux-musl/release/codecraft /out/codecraft-linux-x86_64 \
     && cp target/aarch64-unknown-linux-musl/release/codecraft /out/codecraft-linux-aarch64
 
+# --- Docker CLI (static client; the daemon is the host's via the mounted socket) --
+FROM toolchain AS dockercli
+ARG TARGETARCH
+ARG DOCKER_VERSION=29.9.3
+RUN case "$TARGETARCH" in amd64) arch=x86_64 ;; arm64) arch=aarch64 ;; *) echo "unsupported $TARGETARCH" >&2; exit 1 ;; esac \
+    && curl -fsSL "https://download.docker.com/linux/static/stable/$arch/docker-$DOCKER_VERSION.tgz" | tar -xz -C /tmp docker/docker \
+    && install -m 0755 /tmp/docker/docker /docker
+
 # --- runtime ------------------------------------------------------------------
 # Node is included so agents can work on JS projects and run npx-based MCP servers.
 FROM node:22-bookworm-slim
@@ -85,6 +93,8 @@ RUN npm install -g typescript@5 typescript-language-server pyright && npm cache
 COPY --from=server /codecraft /usr/local/bin/codecraft
 COPY --from=web /web/dist /app/web
 COPY --from=cli /out /app/cli
+# Used with a mounted /var/run/docker.sock (see docs/MOVING.md, --docker-sock).
+COPY --from=dockercli /docker /usr/local/bin/docker
 COPY server/ghidra /app/ghidra
 COPY server/resources /app/resources
 # The agent's home lives on the data volume: tools it installs in user space

README.md

@@ -11,7 +11,7 @@
 </p>
 
 <p align="center">
-  <img alt="Version v202611-00005" src="https://img.shields.io/badge/version-v202611--00005-d69e2e?style=flat-square">
+  <img alt="Version v202611-00007" src="https://img.shields.io/badge/version-v202611--00007-d69e2e?style=flat-square">
   <img alt="License: MIT" src="https://img.shields.io/badge/license-MIT-689d6a?style=flat-square">
   <img alt="Rust backend" src="https://img.shields.io/badge/backend-Rust-b7410e?style=flat-square&logo=rust">
   <img alt="React PWA" src="https://img.shields.io/badge/frontend-React%20PWA-2d81ff?style=flat-square&logo=react">
@@ -103,6 +103,10 @@ copies the workspace, stops CodeCraft, streams the data volume to `<dir>/data`,
 `docker-compose.yml` (and a `start.sh` for hosts without Compose) with bind mounts, starts the
 published image there and runs `scripts/smoke-test.sh`. The old container is only stopped, so
 you can roll back. `--help` lists the options (owner, port, bind address, dry run).
+`--tools-dir /mnt/disks/ssd/codecraft` puts everything agents install or download – the agent
+home (`/data/home`: toolchains, pip/npm/cargo packages, caches), the users' homes and `/tmp` – on
+a separate disk; the data folder keeps only the database, configuration and backups. Step by step,
+with rollback and troubleshooting: [docs/MOVING.md](docs/MOVING.md).
 
 ## How it works
 
@@ -209,6 +213,9 @@ shared secret or a Gitea/GitHub HMAC signature, and its JSON payload is handed t
 
 ## Integrations
 
+- **Telegram:** connect your own bot (*Settings → Account → Telegram*) and use CodeCraft from
+  Telegram – sessions, modes, models, approvals with buttons, photos and files; `/help` in the
+  bot lists the commands.
 - **API tokens** (*Settings → Account*) with full or read-only scope for scripts and the terminal client.
 - **Prometheus:** set `server.metrics_token` and scrape `GET /metrics` with
   `Authorization: Bearer <token>` – requests, errors, tokens, cost and time per model, provider

docs/MOVING.md

@@ -0,0 +1,143 @@
+# Moving CodeCraft to another host (Unraid)
+
+This guide moves a running installation – Docker volume `codecraft-data` on the current host –
+to another Docker host that keeps its data in **bind mounts**, for example Unraid. The script
+`scripts/move-server.sh` does the work; this page explains what goes where, what to prepare and
+how to check and roll back.
+
+## What lives where
+
+| Part | Contents | Today (Docker volume) | On the new host |
+|---|---|---|---|
+| **Data** | database `codecraft.db`, `config.toml` (incl. API keys), daily backups | `codecraft-data:/data` | `<dir>/data` → `/data` |
+| **Agent installs** | the agent's home `/data/home`: toolchains (Rust, …), pip/pipx/npm/cargo packages, downloaded programs (`~/opt`), caches | inside the volume | `<tools-dir>/home` → `/data/home` |
+| **Users' homes** | homes of sandboxed users `/data/users/<name>/home` (their installs, git settings) | inside the volume | `<tools-dir>/users` → `/data/users` |
+| **Temporary files** | `/tmp` of agents, builds, downloads | container layer | `<tools-dir>/tmp` → `/tmp` |
+| **Workspace** | your projects | `WORKSPACE` from `.env` | `--workspace` path → `/workspace` |
+| **Image** | the program | registry image | registry image |
+
+The image never changes: the container runs without root (no `apt`, no `sudo`), so agents can
+only install into their home directory. With `--tools-dir` everything agents install or
+download lands on a separate disk (e.g. an SSD); the data folder keeps only the database,
+configuration and backups. If the tools disk is lost, agents simply install what they need
+again – nothing important is on it. (Backups contain the database and configuration, not the
+agent home.)
+
+Sizes on 2026-11-09 for orientation: agent home 2.2 GB (of which downloaded programs 1.3 GB,
+Rust toolchain 0.8 GB), database/config/backups about 1.2 GB.
+
+## Before you start
+
+On the **new host** (Unraid):
+
+- **SSH with a key** from the current host, without a password prompt. Check from the current
+  host: `ssh root@unraid true` (no output, no question = fine).
+- **docker**, **rsync** and **tar** available (all three are on Unraid by default).
+- **Docker Compose** is optional: without it the script starts the container with the generated
+  `start.sh` (Unraid: the "Docker Compose Manager" plugin adds Compose).
+- The **tools disk** is mounted, e.g. an SSD via Unassigned Devices under `/mnt/disks/ssd`.
+- **Registry access**: the image comes from `git.familie-zink.org/zink/codecraft`. If the pull is
+  refused, run `docker login git.familie-zink.org` on Unraid once.
+
+Decide:
+
+| Question | Option | Example |
+|---|---|---|
+| Folder for data, compose file and scripts | target `root@host:/dir` | `/mnt/user/appdata/codecraft` |
+| Folder for agent installs and `/tmp` | `--tools-dir` | `/mnt/disks/ssd/codecraft` |
+| Your projects | `--workspace` | `/mnt/user/projects` |
+| Owner of files (the container runs as this user) | `--uid`, `--gid` | Unraid default `99`/`100` (nobody:users); today `1000`/`1000` |
+| Docker of the new host for agents and terminals | `--docker-sock` | bind mount of `/var/run/docker.sock` plus its group (see below) |
+| Port and address | `--port`, `--bind` | `8787`; `0.0.0.0` if the reverse proxy runs on another machine or in another Docker network, `127.0.0.1` if it runs on the same host with host networking |
+
+## Moving
+
+On the **current host**, in the repository folder (where `.env` is):
+
+```sh
+# 1. Look at the steps first (changes nothing)
+scripts/move-server.sh root@unraid:/mnt/user/appdata/codecraft \
+  --workspace /mnt/user/projects --tools-dir /mnt/disks/ssd/codecraft \
+  --uid 99 --gid 100 --bind 0.0.0.0 --docker-sock --dry-run
+
+# 2. Move (same command without --dry-run)
+scripts/move-server.sh root@unraid:/mnt/user/appdata/codecraft \
+  --workspace /mnt/user/projects --tools-dir /mnt/disks/ssd/codecraft \
+  --uid 99 --gid 100 --bind 0.0.0.0 --docker-sock
+```
+
+What happens:
+
+1. **Checks**: docker/tar/rsync on the target; the target data folder (and `home/`, `users/` in
+   the tools folder) must be empty – the script never overwrites.
+2. **Workspace, first pass** with rsync while CodeCraft keeps running (the long part).
+3. **Stops CodeCraft here.** Running agents and terminals end. The container is only stopped,
+   not removed.
+4. **Data volume** is streamed with tar over SSH into `<dir>/data`; with `--tools-dir` the agent
+   home and the users' homes are then moved to the tools folder.
+5. **Workspace, second pass**: only what changed since step 2.
+6. **Owner** of data, tools and workspace set to `--uid:--gid` (needs root on the target).
+7. **Writes** `<dir>/docker-compose.yml`, `<dir>/.env` (your `.env` with the new paths, port and
+   owner), `<dir>/start.sh` (the same as `docker run`, for hosts without Compose) and
+   `<dir>/docker/seccomp-sandbox.json`.
+8. **Starts** the image `CODECRAFT_VERSION` from `.env` (Compose, else `start.sh`).
+9. **Smoke test** on the target (`scripts/smoke-test.sh`): health, version, web UI, terminal
+   client download.
+
+Afterwards:
+
+- **Reverse proxy**: point the proxy host for `code.familie-zink.org` (Nginx Proxy Manager) to the
+  new host and port; keep WebSocket support on.
+- **Check in the web UI**: sign in, open a chat, check the status LEDs (providers, MCP servers),
+  let an agent run `git status` in a project (credentials, identity), open a terminal tab.
+  Models or MCP servers on the old host's LAN address or `localhost` need new addresses.
+- **Clean up** on the old host when everything works: `docker rm codecraft`, later
+  `docker volume rm codecraft-data`.
+
+## Docker socket (`--docker-sock`)
+
+With `--docker-sock` the container gets the new host's Docker: `/var/run/docker.sock` is bind
+mounted and the container joins the socket's group (the script reads its group id on the
+target). This is **root-equivalent access to the host**: whoever can start containers can mount
+any host folder.
+
+- Admins' agents and terminals (they run without sandbox) can use it.
+- Sandboxed users do not see the socket (the sandbox does not mount `/var/run`).
+- `docker` and `docker-compose` commands always need an approval, except in *bypass* mode.
+- The image contains the Docker CLI (static client); the daemon is the host's.
+
+The current host has the same setup in its local `docker-compose.override.yml` (group 997).
+
+## Rollback
+
+Nothing is deleted on the old host. On Unraid stop the new container (`docker stop codecraft`),
+on the old host start the old one (`docker start codecraft`) and point the reverse proxy back.
+Changes made on the new host in the meantime (chats, settings) are not in the old installation.
+
+## Running it on the new host
+
+In `<dir>` on Unraid:
+
+| Task | With Compose | Without Compose |
+|---|---|---|
+| Update | set `CODECRAFT_VERSION` in `.env`, then `docker compose pull && docker compose up -d` | set `CODECRAFT_VERSION` in `.env`, `docker pull …:<version>`, then `./start.sh` |
+| Restart | `docker compose restart` | `docker restart codecraft` |
+| Logs | `docker compose logs -f` | `docker logs -f codecraft` |
+| Check | `bash smoke-test.sh http://127.0.0.1:8787 <version>` | same |
+
+Releases are published from the development machine with `scripts/release.sh`; the new host
+only pulls them.
+
+## Troubleshooting
+
+- **`data on the target is not empty`**: an earlier attempt left files. Move `<dir>/data` (and
+  `home/`, `users/` in the tools folder) away and run the script again.
+- **Image pull refused**: `docker login git.familie-zink.org` on the target, then
+  `cd <dir> && docker compose up -d` (or `./start.sh`).
+- **Smoke test fails**: `docker logs codecraft` on the target. The old installation is stopped but
+  intact – `docker start codecraft` on the old host brings it back.
+- **`could not chown`**: the SSH user is not root; set the owner by hand
+  (`chown -R 99:100 <dir>/data <tools-dir> <workspace>`).
+- **Permission denied inside the container**: data, tools and workspace must belong to the
+  `UID`/`GID` in `<dir>/.env`.
+- **Port already in use**: change `PORT` in `<dir>/.env` and start again.

scripts/move-server.sh

@@ -23,7 +23,14 @@
 #
 # Options:
 #   --workspace DIR      project folder on the target (default <dir>/workspace)
+#   --tools-dir DIR      separate disk for what agents install and download: the agent home
+#                        (/data/home), the users' homes (/data/users) and /tmp are bind
+#                        mounts from DIR/home, DIR/users and DIR/tmp (default: inside <dir>/data)
 #   --uid N --gid N      owner of data and workspace on the target (default: UID/GID of .env)
+#   --docker-sock        give the container the target's Docker (bind mount of
+#                        /var/run/docker.sock plus its group). Root-equivalent access to the
+#                        target host for admins' agents and terminals; sandboxed users do not
+#                        see it, and docker commands still need approval
 #   --port N             host port on the target (default: PORT of .env)
 #   --bind ADDR          bind address on the target (default: BIND_ADDR of .env or 127.0.0.1)
 #   --name NAME          container name on the target (default codecraft)
@@ -34,11 +41,13 @@
 #   --dry-run            show the steps only
 set -euo pipefail
 
-target="" ws_target="" uid="" gid="" port="" bind="" name="codecraft"
+target="" ws_target="" tools_dir="" docker_sock=0 uid="" gid="" port="" bind="" name="codecraft"
 src_container="codecraft" src_volume="codecraft-data" ws_source="" stop=1 dry=0
 while (($#)); do
   case "$1" in
     --workspace) ws_target="$2"; shift ;;
+    --tools-dir) tools_dir="$2"; shift ;;
+    --docker-sock) docker_sock=1 ;;
     --uid) uid="$2"; shift ;;
     --gid) gid="$2"; shift ;;
     --port) port="$2"; shift ;;
@@ -97,10 +106,15 @@ step "Checking source and target"
 docker volume inspect "$src_volume" >/dev/null || fail "volume $src_volume not found"
 [[ -d "$ws_source" ]] || fail "workspace $ws_source not found"
 echo "source: container $src_container, volume $src_volume, workspace $ws_source"
-echo "target: $host:$dir (data in $dir/data, workspace $ws_target), uid $uid:$gid, ${bind}:${port}, image $image_repo:$version"
+tools_dir="${tools_dir%/}"
+echo "target: $host:$dir (data in $dir/data, workspace $ws_target${tools_dir:+, agent installs and /tmp in $tools_dir}), uid $uid:$gid, ${bind}:${port}, image $image_repo:$version"
 remote "command -v docker >/dev/null && command -v tar >/dev/null && command -v rsync >/dev/null" || fail "the target needs docker, tar and rsync"
 remote "mkdir -p $(q "$dir/data") $(q "$dir/docker") $(q "$ws_target") && test -z \"\$(ls -A $(q "$dir/data"))\"" \
   || fail "$dir/data on the target is not empty – refusing to overwrite (move it away first)"
+if [[ -n "$tools_dir" ]]; then
+  remote "mkdir -p $(q "$tools_dir") && test ! -e $(q "$tools_dir/home") && test ! -e $(q "$tools_dir/users")" \
+    || fail "$tools_dir already has home/ or users/ – refusing to overwrite (move them away first)"
+fi
 
 step "Copying the workspace (first pass, CodeCraft keeps running)"
 sync_workspace
@@ -118,9 +132,29 @@ else
     | remote "tar -C $(q "$dir/data") -xpf -"
 fi
 
+if [[ -n "$tools_dir" ]]; then
+  step "Moving the agent home and the users' homes to $tools_dir"
+  remote "mkdir -p $(q "$tools_dir/tmp") $(q "$dir/data/home") $(q "$dir/data/users") \
+    && mv $(q "$dir/data/home") $(q "$tools_dir/home") && mv $(q "$dir/data/users") $(q "$tools_dir/users") \
+    && mkdir -p $(q "$dir/data/home") $(q "$dir/data/users") && chmod 1777 $(q "$tools_dir/tmp")"
+fi
+
 step "Copying the workspace (changes since the first pass)"
 sync_workspace
-remote "chown -R $uid:$gid $(q "$dir/data") $(q "$ws_target") 2>/dev/null || echo 'note: could not chown (not root) – check file owners'"
+remote "chown -R $uid:$gid $(q "$dir/data") $(q "$ws_target") ${tools_dir:+$(q "$tools_dir/home") $(q "$tools_dir/users")} 2>/dev/null || echo 'note: could not chown (not root) – check file owners'"
+# Bind mounts for the tools disk and the Docker socket (compose lines and docker run flags).
+tool_mounts="" tool_flags="" sock_compose="" sock_flags=""
+if (( docker_sock )); then
+  if (( dry )); then sock_gid=0; else
+    sock_gid=$(remote "stat -c %g /var/run/docker.sock") || fail "no /var/run/docker.sock on the target"
+  fi
+  sock_compose=$(printf '      - /var/run/docker.sock:/var/run/docker.sock\n    group_add:\n      - "%s"\n' "$sock_gid")
+  sock_flags="-v /var/run/docker.sock:/var/run/docker.sock --group-add $sock_gid "
+fi
+if [[ -n "$tools_dir" ]]; then
+  tool_mounts=$(printf '      - %s/home:/data/home\n      - %s/users:/data/users\n      - %s/tmp:/tmp\n' "$tools_dir" "$tools_dir" "$tools_dir")
+  tool_flags="-v $tools_dir/home:/data/home -v $tools_dir/users:/data/users -v $tools_dir/tmp:/tmp "
+fi
 
 step "Writing docker-compose.yml, .env and start.sh"
 put_file "$dir/docker/seccomp-sandbox.json" <docker/seccomp-sandbox.json
@@ -145,6 +179,8 @@ services:
     volumes:
       - $dir/data:/data
       - \${WORKSPACE}:/workspace
+$tool_mounts
+$sock_compose
     extra_hosts:
       - "host.docker.internal:host-gateway"
     shm_size: "1gb"
@@ -163,7 +199,7 @@ val() { sed -n "s/^\$1=//p" .env | tail -1; }
 docker rm -f $name >/dev/null 2>&1 || true
 exec docker run -d --name $name --restart unless-stopped --user "\$(val UID):\$(val GID)" \\
   -p "\$(val BIND_ADDR):\$(val PORT):8787" --env-file .env \\
-  -v $dir/data:/data -v "\$(val WORKSPACE):/workspace" \\
+  -v $dir/data:/data -v "\$(val WORKSPACE):/workspace" $tool_flags$sock_flags\\
   --add-host host.docker.internal:host-gateway --shm-size 1g \\
   --security-opt seccomp=$dir/docker/seccomp-sandbox.json \\
   --security-opt apparmor=unconfined --security-opt systempaths=unconfined \\

server/Cargo.toml

@@ -28,7 +28,7 @@ shell-words = "1"
 libc = "0.2"
 ignore = "0.4"
 regex = "1"
-reqwest = { version = "0.13", default-features = false, features = ["json", "stream", "rustls", "form"] }
+reqwest = { version = "0.13", default-features = false, features = ["json", "stream", "rustls", "form", "multipart"] }
 rusqlite = { version = "0.41", features = ["bundled"] }
 serde = { version = "1", features = ["derive"] }
 serde_json = "1"

server/src/api.rs

@@ -138,6 +138,7 @@ pub fn router(app: AppState) -> Router {
         .route("/account/prefs", get(crate::cli_api::get_prefs).put(crate::cli_api::put_prefs))
         .route("/account/memory", get(crate::cli_api::get_memory).put(crate::cli_api::put_memory))
         .route("/account/modes", get(get_personal_modes).put(put_personal_modes))
+        .route("/account/telegram", get(get_telegram).put(put_telegram).delete(delete_telegram))
         .route("/templates", get(crate::cli_api::get_templates))
         .route("/templates/{scope}", put(crate::cli_api::put_templates))
         .route("/cli/config", get(crate::cli_api::config))
@@ -327,6 +328,43 @@ async fn agents(Extension(user): Extension<User>, State(app): State<AppState>, Q
     Json(json!(prompts::agents_for(&app, &root, &cwd, Some(&user.id))))
 }
 
+/// Telegram bot of the user: bot name, pairing state, errors (the token stays hidden).
+async fn get_telegram(Extension(user): Extension<User>, State(app): State<AppState>) -> ApiResult {
+    let s = crate::telegram::load(&app, &user.id);
+    let (running, error) = crate::telegram::status(&user.id);
+    Ok(Json(json!({
+        "configured": !s.token.is_empty(),
+        "bot": s.bot,
+        "paired": s.chat_id.is_some(),
+        "pair_code": if s.chat_id.is_none() { s.pair_code } else { String::new() },
+        "running": running,
+        "error": error,
+    })))
+}
+
+#[derive(Deserialize)]
+struct TelegramBody {
+    token: String,
+}
+
+/// Connect a bot: check the token, start polling and create a pairing code.
+async fn put_telegram(Extension(user): Extension<User>, State(app): State<AppState>, Json(b): Json<TelegramBody>) -> ApiResult {
+    let token = b.token.trim().to_string();
+    let bot = crate::telegram::check_token(&token).await.map_err(|e| ApiError(StatusCode::BAD_REQUEST, format!("token rejected by Telegram: {e:#}")))?;
+    let s = crate::telegram::Settings { token, bot, pair_code: crate::telegram::new_pair_code(), ..Default::default() };
+    app.store.set_user_setting(&user.id, "telegram", &s)?;
+    crate::telegram::restart(&app, &user.id);
+    app.audit(Some(&user), "telegram.connect", &s.bot, "");
+    get_telegram(Extension(user), State(app)).await
+}
+
+async fn delete_telegram(Extension(user): Extension<User>, State(app): State<AppState>) -> ApiResult {
+    crate::telegram::stop(&user.id);
+    app.store.set_user_setting(&user.id, "telegram", &crate::telegram::Settings::default())?;
+    app.audit(Some(&user), "telegram.disconnect", "", "");
+    Ok(Json(json!({"ok": true})))
+}
+
 /// The user's personal modes (name -> settings).
 async fn get_personal_modes(Extension(user): Extension<User>, State(app): State<AppState>) -> ApiResult {
     Ok(Json(json!(prompts::personal_modes(&app, &user.id))))
@@ -858,41 +896,48 @@ fn set_turn_effort(app: &App, id: &str, effort: Option<&str>) -> Result<(), ApiE
 
 async fn prompt(Extension(user): Extension<User>, State(app): State<AppState>, Path(id): Path<String>, Json(body): Json<PromptBody>) -> ApiResult {
     let session = writable(&app, &user, &id)?;
-    let text = body.text.trim().to_string();
+    send_prompt(&app, session, body.text, body.images, body.effort.as_deref()).map(Json)
+}
+
+/// Handle a user message for a session (web UI and Telegram): built-in commands,
+/// slash command expansion, then start or queue the turn.
+pub(crate) fn send_prompt(app: &AppState, session: crate::types::Session, text: String, images: Vec<ImageData>, effort: Option<&str>) -> Result<Value, ApiError> {
+    let id = session.id.clone();
+    let text = text.trim().to_string();
     // Built-in commands.
     if text == "/compact" {
         if app.is_busy(&id) {
             return Err(ApiError(StatusCode::CONFLICT, "session is busy".into()));
         }
         tokio::spawn(do_compact(app.clone(), id.clone()));
-        return Ok(Json(json!({"ok": true, "command": "compact"})));
+        return Ok(json!({"ok": true, "command": "compact"}));
     }
     let cwd = PathBuf::from(&session.cwd);
     let root = app.root_for(session.owner_id.as_deref(), &cwd);
     let mut prompt_text = text.clone();
     if text.starts_with('/') {
-        if let Some((expanded, agent, clean_room)) = prompts::expand_command(&app, &root, &cwd, &text) {
+        if let Some((expanded, agent, clean_room)) = prompts::expand_command(app, &root, &cwd, &text) {
             prompt_text = expanded;
             if agent.is_some() || clean_room.is_some() {
-                let mut s = session.clone();
-                if let Some(a) = agent {
-                    s.agent = a;
-                }
-                if clean_room.is_some() {
-                    s.clean_room = clean_room;
-                }
-                app.save_session(s)?;
+                app.update_session(&id, |s| {
+                    if let Some(a) = agent {
+                        s.agent = a;
+                    }
+                    if clean_room.is_some() {
+                        s.clean_room = clean_room;
+                    }
+                })?;
             }
         }
     }
-    set_turn_effort(&app, &id, body.effort.as_deref())?;
+    set_turn_effort(app, &id, effort)?;
     let mut parts = vec![Part::Text { text: prompt_text }];
-    for image in body.images {
+    for image in images {
         parts.push(Part::Image { image });
     }
     match runner::submit(app.clone(), id, parts, "user")? {
-        runner::Submitted::Started => Ok(Json(json!({"ok": true}))),
-        runner::Submitted::Queued(item) => Ok(Json(json!({"ok": true, "queued": item}))),
+        runner::Submitted::Started => Ok(json!({"ok": true})),
+        runner::Submitted::Queued(item) => Ok(json!({"ok": true, "queued": item})),
     }
 }
 
@@ -916,7 +961,7 @@ async fn flush_queue(Extension(user): Extension<User>, State(app): State<AppStat
     Ok(Json(json!({"started": runner::flush_queue(app.clone(), &id)?})))
 }
 
-async fn do_compact(app: AppState, id: String) {
+pub(crate) async fn do_compact(app: AppState, id: String) {
     let cancel = tokio_util::sync::CancellationToken::new();
     app.with_runtime(&id, |r| r.cancel = Some(cancel.clone()));
     app.emit(Event::SessionStatus { session_id: id.clone(), busy: true });

server/src/main.rs

@@ -30,6 +30,7 @@ mod runner;
 mod sandbox;
 mod shells;
 mod store;
+mod telegram;
 mod terminal;
 mod tools;
 mod types;
@@ -73,6 +74,7 @@ async fn main() -> Result<()> {
     health::spawn_checker(app.clone());
     digest::spawn(app.clone());
     budget::spawn_alerts(app.clone());
+    telegram::spawn(app.clone());
     backup::spawn(app.clone());
     let router = api::router(app);
     let listener = tokio::net::TcpListener::bind(&bind).await?;

web/src/components/settings/Account.tsx

@@ -6,6 +6,7 @@ import { GitAccount } from "./GitAccount";
 import { ApiTokens } from "./ApiTokens";
 import { CliInstall } from "./CliInstall";
 import { Memory } from "./Memory";
+import { Telegram } from "./Telegram";
 
 export function Account() {
   const t = useT();
@@ -51,6 +52,7 @@ export function Account() {
     <Memory />
     <GitAccount />
     <CliInstall />
+    <Telegram />
     <ApiTokens />
     </>
   );

web/src/components/settings/Telegram.tsx

@@ -0,0 +1,94 @@
+import { useState } from "react";
+import { del, put } from "../../api";
+import { confirmDialog, toast } from "../../dialogs";
+import { useFetch } from "../../hooks";
+import { useT } from "../../i18n";
+import { Icon } from "../Icon";
+
+interface Status {
+  configured: boolean;
+  bot: string;
+  paired: boolean;
+  pair_code: string;
+  running: boolean;
+  error?: string | null;
+}
+
+/** Connect a personal Telegram bot as a chat interface. */
+export function Telegram() {
+  const t = useT();
+  const { data, reload } = useFetch<Status>("/account/telegram");
+  const [token, setToken] = useState("");
+  const [busy, setBusy] = useState(false);
+
+  const connect = async () => {
+    setBusy(true);
+    try {
+      await put("/account/telegram", { token: token.trim() });
+      setToken("");
+      reload();
+    } catch (e) {
+      toast(e);
+    } finally {
+      setBusy(false);
+    }
+  };
+  const disconnect = async () => {
+    if (!(await confirmDialog({ message: t("telegram.disconnectConfirm"), danger: true, confirmLabel: t("telegram.disconnect") }))) return;
+    await del("/account/telegram").catch((e) => toast(e));
+    reload();
+  };
+
+  return (
+    <div className="panel">
+      <h3>
+        <Icon name="share" /> {t("telegram.title")}
+        <span className="spacer" />
+        {data?.configured && (
+          <button className="small ghost danger" onClick={disconnect}>
+            {t("telegram.disconnect")}
+          </button>
+        )}
+      </h3>
+      <div className="hint" style={{ marginBottom: 10 }}>
+        {t("telegram.intro")}
+      </div>
+      {data && !data.configured && (
+        <>
+          <ol className="cli-steps">
+            <li>{t("telegram.step1")}</li>
+            <li>{t("telegram.step2")}</li>
+          </ol>
+          <div className="row" style={{ gap: 6 }}>
+            <input type="password" value={token} placeholder="123456789:AA…" autoComplete="off" onChange={(e) => setToken(e.target.value)} style={{ flex: 1 }} />
+            <button className="primary" onClick={() => void connect()} disabled={busy || !token.trim()}>
+              {busy ? t("telegram.checking") : t("telegram.connect")}
+            </button>
+          </div>
+        </>
+      )}
+      {data?.configured && (
+        <>
+          <div className="row" style={{ gap: 8, alignItems: "center", flexWrap: "wrap" }}>
+            <span className="chip">@{data.bot}</span>
+            <span className={`chip ${data.running && !data.error ? "ok" : "warn"}`}>{data.running && !data.error ? t("telegram.running") : t("telegram.stopped")}</span>
+            {data.paired && <span className="chip ok">{t("telegram.paired")}</span>}
+          </div>
+          {!data.paired && data.pair_code && (
+            <div className="notice" style={{ marginTop: 10 }}>
+              {t("telegram.pairHint", { bot: data.bot })}{" "}
+              <a href={`https://t.me/${data.bot}?start=${data.pair_code}`} target="_blank" rel="noreferrer">
+                t.me/{data.bot}
+              </a>
+              <div style={{ marginTop: 6 }}>
+                <code className="token-value">/start {data.pair_code}</code>
+              </div>
+            </div>
+          )}
+          {data.error && <div className="notice err" style={{ marginTop: 10 }}>{data.error}</div>}
+          {data.paired && <div className="hint" style={{ marginTop: 10 }}>{t("telegram.commands")}</div>}
+        </>
+      )}
+    </div>
+  );
+}

web/src/i18n/de.ts

@@ -405,6 +405,19 @@ export const de: Messages = {
   "tokens.title": "API-Tokens",
   "tokens.hint": "Für den Kommandozeilen-Client (codecraft) und Skripte: als „Authorization: Bearer <token>“ mitschicken. Nur-Lesen-Tokens können nur lesen (keine Prompts, keine Änderungen).",
   "tokens.copyNow": "Token jetzt kopieren – er wird nur einmal angezeigt:",
+  "telegram.title": "Telegram",
+  "telegram.intro": "CodeCraft über Telegram mit deinem eigenen Bot nutzen: Nachrichten gehen an deine aktuelle Sitzung, Befehle wechseln Sitzungen, Modi und Modelle, Freigaben und Rückfragen kommen mit Knöpfen. Nur dein Telegram-Konto kann den Bot nutzen.",
+  "telegram.step1": "In Telegram @BotFather öffnen, /newbot senden und den Schritten folgen.",
+  "telegram.step2": "Das Token von BotFather hier einfügen.",
+  "telegram.connect": "Verbinden",
+  "telegram.checking": "Prüfe…",
+  "telegram.running": "läuft",
+  "telegram.stopped": "läuft nicht",
+  "telegram.paired": "gekoppelt",
+  "telegram.pairHint": "Letzter Schritt: deinen Bot öffnen und den Code senden (oder den Link antippen) – danach bedient er nur noch dein Telegram-Konto:",
+  "telegram.commands": "In Telegram: /help zeigt die Befehle – /sessions, /new, /mode, /model, /stop, /status, /costs …",
+  "telegram.disconnect": "Trennen",
+  "telegram.disconnectConfirm": "Telegram-Bot trennen? Er antwortet dann nicht mehr; mit seinem Token kannst du ihn wieder verbinden.",
   "tokens.copy": "Kopieren",
   "tokens.copied": "Kopiert",
   "tokens.name": "Name",

web/src/i18n/en.ts

@@ -404,6 +404,19 @@ export const en = {
   "tokens.title": "API tokens",
   "tokens.hint": "For the command line client (codecraft) and scripts: send as “Authorization: Bearer <token>”. Read-only tokens can only read (no prompts, no changes).",
   "tokens.copyNow": "Copy the token now – it is shown only once:",
+  "telegram.title": "Telegram",
+  "telegram.intro": "Use CodeCraft from Telegram with your own bot: messages go to your current session, commands switch sessions, modes and models, approvals and questions come with buttons. Only your Telegram account can use the bot.",
+  "telegram.step1": "In Telegram, open @BotFather, send /newbot and follow the steps.",
+  "telegram.step2": "Paste the token BotFather gives you here.",
+  "telegram.connect": "Connect",
+  "telegram.checking": "Checking…",
+  "telegram.running": "running",
+  "telegram.stopped": "not running",
+  "telegram.paired": "paired",
+  "telegram.pairHint": "Last step: open your bot and send the code (or tap the link) – from then on only your Telegram account is served:",
+  "telegram.commands": "In Telegram: /help lists the commands – /sessions, /new, /mode, /model, /stop, /status, /costs …",
+  "telegram.disconnect": "Disconnect",
+  "telegram.disconnectConfirm": "Disconnect the Telegram bot? It stops answering; you can connect it again with its token.",
   "tokens.copy": "Copy",
   "tokens.copied": "Copied",
   "tokens.name": "Name",

web/src/whatsnew.ts

@@ -3,6 +3,14 @@
  * under "unreleased"; scripts/release.sh renames the key to the release version.
  */
 export const WHATS_NEW: Record<string, { en: string[]; de: string[] }> = {
+  "202611-00007": {
+    en: ["Telegram: connect your own bot under Settings → Account → Telegram and chat with CodeCraft from your phone"],
+    de: ["Telegram: eigenen Bot unter Einstellungen → Konto → Telegram verbinden und von unterwegs mit CodeCraft chatten"],
+  },
+  "202611-00006": {
+    en: ["Docker CLI in the image – with the host's Docker socket mounted, admins' agents and terminals can use Docker"],
+    de: ["Docker-CLI im Image – mit eingehängtem Docker-Socket können Admin-Agents und -Terminals Docker nutzen"],
+  },
   "202611-00005": {
     en: ["Agents commit with the git name and e-mail from your settings – also in repositories with their own identity"],
     de: ["Agents committen mit Git-Name und -E-Mail aus deinen Einstellungen – auch in Repos mit eigener Identität"],