Skip to content

feat: Add /add-telegram skill for Telegram channel support#33

Closed
randomm wants to merge 6 commits intoqwibitai:mainfrom
randomm:main
Closed

feat: Add /add-telegram skill for Telegram channel support#33
randomm wants to merge 6 commits intoqwibitai:mainfrom
randomm:main

Conversation

@randomm
Copy link
Copy Markdown
Contributor

@randomm randomm commented Feb 2, 2026

Summary

Adds the /add-telegram skill to enable Telegram as a communication channel for NanoClaw.

This addresses the RFS (Request for Skills) item in the README for /add-telegram.

What the Skill Does

When a user runs /add-telegram, Claude guides them through adding Telegram support with four modes:

  1. Replace WhatsApp - Telegram becomes the primary channel
  2. Additional Channel - Run both Telegram and WhatsApp simultaneously
  3. Control Channel - Telegram triggers actions, WhatsApp remains primary
  4. Action-Only - Telegram for notifications only (no incoming handler)

Why Telegram is a Good Addition

Aspect Telegram WhatsApp
Official API ✅ Yes ❌ No
Setup time 2 minutes 5+ minutes
Phone required No Yes
ToS ✅ Bots permitted ⚠️ Against ToS

Skill Features

  • Deterministic implementation - Exact code patterns, no research needed
  • Complete setup guide - @Botfather flow, token storage, chat ID discovery
  • Privacy model explained - How registered chats provide access control
  • All gotchas documented - Group privacy mode, rate limits, message prefixes
  • Tested in production - Friction points identified and fixed

Files Added

  • .claude/skills/add-telegram/SKILL.md (700+ lines)

Testing

The skill was tested end-to-end with "Replace WhatsApp" mode. Issues discovered during testing were fixed:

  • dotenv dependency added
  • Chat ID discovery improved
  • Service conflict warnings added
  • Message prefix handling (Telegram bots don't need ASSISTANT_NAME: prefix)
  • Privacy model documentation added

Janni Turunen and others added 6 commits February 2, 2026 11:06
Adds a skill to integrate Parallel AI MCP servers for advanced web research.

What it adds:
- Quick Search MCP for fast web lookups (free to use)
- Task MCP for deep research with scheduler-based polling
- Non-blocking design using NanoClaw's scheduler

Usage: Run /add-parallel to add this integration to your fork

Follows NanoClaw philosophy:
- Skills over features - integration is optional via skill
- Keeps base codebase lean
- Users run /add-parallel on their fork for clean code

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Applied the /add-parallel skill to enable Parallel AI on this fork.

Changes:
- Added PARALLEL_API_KEY to container environment variables
- Configured Parallel AI MCP servers (Search & Task)
- Added Web Research Tools section to CLAUDE.md
- Updated README with Parallel AI integration docs
- Improved container build script with Docker fallback
- Added ASSISTANT_NAME configuration support

This fork now has Parallel AI integrated with:
- Quick Search for fast web lookups
- Deep Research with scheduler-based polling
- Non-blocking design using NanoClaw's scheduler

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
… testing

- Add dotenv dependency installation and import guidance
- Add bot token validation step using getMe API
- Clarify chat ID discovery process with correct index (.result[-1])
- Add service conflict warning for existing launchctl service
- Explain message prefix handling (Telegram bots send as themselves)
- Add comprehensive privacy model explanation and registered_groups.json format
- Include example of multiple registered chats
- Add bot naming guidance (anthropomorphic names preferred)
- Improve clarity of all prerequisites section
@randomm randomm requested a review from gavrielc as a code owner February 2, 2026 12:54
@randomm
Copy link
Copy Markdown
Contributor Author

randomm commented Feb 2, 2026

Closing - PR contained unrelated fork changes. Will resubmit with only the telegram skill.

@randomm randomm closed this Feb 2, 2026
jbaruch pushed a commit to jbaruch/nanoclaw-public that referenced this pull request Apr 9, 2026
- SOUL.md injected via systemPrompt.append (survives compaction)
- Untrusted prompts tagged with <untrusted-input> origin markers
- Circuit breaker pauses groups after 5 consecutive failures (30min cooldown)
- settings.json writes skipped when content unchanged (prompt cache optimization)
- New rules: verification protocol, tone matching, query size limits, post-compaction trust

Closes #27, #28, qwibitai#29, qwibitai#31, qwibitai#32, qwibitai#33, qwibitai#34

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
kenansun-dev-bot Bot referenced this pull request in kenansun-dev/nanoclaw-github-copilot Apr 12, 2026
… empty (#33)

If ~/.nanoclaw/skills/ is empty, use container/skills/ from the repo.
This ensures host mode has access to built-in skills (teams-formatting etc).

344/344 tests pass.

Co-authored-by: Kenan VM Claw <vm-claw@kenan.dev>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant