feat: claude code runner (#9)

readme.md

@@ -2,22 +2,26 @@
 
 🐙 *he just wants to help-pi*
 
-telegram bot for [codex](https://github.com/openai/codex). runs `codex exec --json`, streams progress, and supports resumable sessions.
+telegram bridge for codex and claude code. runs the agent cli, streams progress, and supports resumable sessions.
 
 ## features
 
-stateless resume via `codex resume <token>` lines in chat.
+stateless resume, continue a thread in the chat or pick up in the terminal.
 
-edits a single progress message while codex runs (commands, tools, notes, file changes, elapsed time).
+progress updates while agent runs (commands, tools, notes, file changes, elapsed time).
 
-renders markdown to telegram entities.
+robust markdown rendering of output with a lot of quality of life tweaks.
 
-runs in parallel across threads and queues per thread to keep codex history sane.
+parallel runs across threads, per thread queue support.
+
+`/cancel` a running task.
 
 ## requirements
 
 - `uv` for installation (`curl -LsSf https://astral.sh/uv/install.sh | sh`)
-- `codex` on PATH (`npm install -g @openai/codex` or `brew install codex`)
+- at least one engine installed:
+  - `codex` on PATH (`npm install -g @openai/codex` or `brew install codex`)
+  - `claude` on PATH (`npm install -g @anthropic-ai/claude-code`)
 
 ## install
 
@@ -29,12 +33,11 @@ runs in parallel across threads and queues per thread to keep codex history sane
 1. get `bot_token` from [@BotFather](https://t.me/BotFather)
 2. get `chat_id` from [@myidbot](https://t.me/myidbot)
 3. send `/start` to the bot (telegram won't let it message you first)
-4. run `codex` once interactively in the repo to trust the directory
+4. run your agent cli once interactively in the repo to trust the directory
 
 ## config
 
-takopi reads `.takopi/takopi.toml` in the current repo, otherwise `~/.takopi/takopi.toml`.
-legacy `.codex/takopi.toml` is migrated automatically.
+global config `~/.takopi/takopi.toml`, repo-level config `.takopi/takopi.toml`
 
 ```toml
 bot_token = "123456789:ABCdefGHIjklMNOpqrsTUVwxyz"
@@ -43,6 +46,13 @@ chat_id = 123456789
 [codex]
 # optional: profile from ~/.codex/config.toml
 profile = "takopi"
+
+[claude]
+model = "sonnet"
+allowed_tools = ["Bash", "Read", "Write", "WebSearch"]
+dangerously_skip_permissions = false
+# uses subscription by default, override to use api billing
+use_api_billing = false
 ```
 
 ## usage
@@ -51,27 +61,26 @@ start takopi in the repo you want to work on:
 
 ```sh
 cd ~/dev/your-repo
-takopi
+takopi codex
+# or
+takopi claude
 ```
 
 send a message to the bot.
 
 to continue a thread, reply to a bot message containing a resume line.
+you can also copy it to resume an interactive session in your terminal.
 
 to stop a run, reply to the progress message with `/cancel`.
 
-## cli
+default: progress is silent, final answer is sent as a new message so you receive a notification, progress message is deleted.
 
-default: progress is silent, final answer is sent as a new message (notification), progress message is deleted.
-
-`--no-final-notify` edits the progress message into the final answer (no new notification).
-
-`--debug` enables verbose logs.
+if you prefer no notifications, `--no-final-notify` edits the progress message into the final answer.
 
 ## notes
 
-* private chat only
-* run exactly one instance per bot token
+* private chat only: the bot only responds to the configured `chat_id`
+* run only one takopi instance per bot token: multiple instances will race telegram's `getUpdates` offsets and cause missed updates
 
 ## development