hanxiao/claudecode-telegram: source audit, 2026-10-10
Read from the source at commit c925f90e0b54; the bot was not run (it needs a BotFather token). Each answer cites the file and line that shows it. All 20 bots
In short
Anyone who messages the bot, and anyone who can POST to the webhook URL, gets their text typed into a Claude Code session running with permissions skipped. Do not expose it without adding your own sender check and webhook secret.
What the source shows
| Answer | Where | |
|---|---|---|
| Who can use it | Has no allowlist at all: anyone who finds the bot can use it | bridge.py:162-169 — handle_message reads chat_id and writes it to a file with no sender check; bridge.py:262 — tmux_send(text) |
| With no allowlist set | none exists: no allowlist, chat ID check or webhook secret in bridge.py | bridge.py:113-120 — do_POST handles any update with no secret-token check; bridge.py:279 — HTTPServer(("0.0.0.0", PORT), Handler) |
| Before the agent acts | runs the agent with permission prompts switched off | bridge.py:148 — claude --resume {session_id} --dangerously-skip-permissions; also :158 and :207 |
| Bot token | Read from the environment or a config file; no real token is committed | |