1.8k
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill openclaw/skills --skill tg-channel-manager- _meta.json467 B
- SKILL.md8.7 KB
Overview
This skill is a universal, config-driven content pipeline engine for managing any Telegram channel. It automates news discovery, drafts creation, human approval flow, deduplication, and scheduled publishing. All channel behavior is defined via a JSON config so the same skill works for many channels without code changes.
How this skill works
The pipeline runs three steps: scout (cron) discovers news via SearXNG, applies include/exclude filters and dedup checks, then writes drafts to a workspace queue. A human reviewer approves drafts by changing status to pending. The publisher (cron) reads pending entries, formats posts by rubric and style rules, publishes a single post per run via the Telegram messaging tool, updates the dedup index, and removes the published entry.
When to use it
- You run a Telegram channel that needs automated content discovery and scheduled posting.
- You need a repeatable, config-driven workflow that separates scouting, human approval, and publishing.
- You want automatic deduplication across drafts and published posts.
- You require strict post formatting rules (length, emoji, signature) and rubric rotation.
- You need evergreen article insertion when news is scarce.
Best practices
- Define clear searchQueries, include/exclude filters, and evergreen topics in the config.
- Set sensible limits: maxDraftsPerRun and maxPostsPerDay to avoid flooding the channel.
- Keep postStyle constraints aligned with channel voice and character limits.
- Run deduplication rebuilds periodically if you change channelId or need a full sync.
- Review drafts regularly and mark pending only after light editing to preserve pipeline cadence.
Example use cases
- A tech news channel: auto-scout new posts from SearXNG, draft headlines, human approves and publisher posts with signature.
- A documentation channel: publish official changelog links and short summaries with primary-source preference.
- A niche community channel: use rubric rotation and timezone-aware cron to balance categories.
- An archive/backup channel: deduplicate and index historical posts, rebuild index from Telegram when needed.
FAQ
All settings live in openclaw.json under the skill entry: channel IDs, schedules, rubrics, search queries and postStyle — no hardcoded values.
How does deduplication work?
Before each draft the scout runs dedup-check.py against content-index.json; after publishing the message ID and links are added to the index. You can rebuild the index via the provided script and Telegram search.