Skip to content
Back to skills

Telegram Output Formatting

ASecurity

How to format Telegram output for Gregor — short summaries with emoji titles, bullet-point lists, and full reports committed to plur-space. Applies to every report, briefing, task summary, and intelligence output delivered in Telegram DMs.

  • 4 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added September 26, 2026
ai-agentspythongogitapi

Works with

  • api

Security analysis

A100/100

Scanned September 26, 2026

npx -y skills add datacore-one/datacore --skill telegram-output-formatting --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Telegram Output Formatting?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Telegram Output Formatting
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/datacore-one-telegram-output-formatting/badge)](https://www.skillsdirectory.com/skills/datacore-one-telegram-output-formatting)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: telegram-output-formatting
description: >
  How to format Telegram output for Gregor — short summaries with emoji titles,
  bullet-point lists, and full reports committed to plur-space. Applies to every
  report, briefing, task summary, and intelligence output delivered in Telegram DMs.
triggers:
  - reporting results in Telegram
  - summarizing task output
  - formatting lists or bullet points
  - delivering intelligence briefings
  - "format nicer"
  - "make it shorter"
version: 1.0.0
---

# Telegram Output Formatting

## The Rule

Telegram is the memo. The disk is the dossier.

Chat output = short, punchy, scannable.
Full reports = committed to the space's journal (`<space>/journal/`).

## Format Rules

### Titles
- Use emojis in section titles: ✅ 🔴 🟢 🟡 📋 🔍 💡 ⚠️
- Titles are SHORT — 2-4 words max
- CAPS for status markers when needed (OK, ERROR, DONE)

### Lists
- Use bullet points (•) — NOT hyphens (-)
- One line per item, no nested sub-bullets deeper than one level
- Keep items to a single sentence

### Length
- Telegram messages should fit on one phone screen
- Highlights and takeaways only — "what matters, what's next"
- If the output would exceed ~15 lines, write a full report to disk and
  summarize in chat with a pointer to the file

### Structure
A typical task report:

```
emoji TITLE

• Key finding 1
• Key finding 2
• Key finding 3

Next: what to do about it
```

For multi-item status reports (cron jobs, task lists):

```
📋 Header

🟢 Item 1 — status
🔴 Item 2 — status
🟢 Item 3 — status

Summary line
```

### What goes to disk, not chat
- Full methodology and sources
- Raw data dumps (GitHub API responses, scan logs)
- Multi-paragraph analysis
- Reproduction steps for technical issues
- Anything that would make the message scroll past one screen

Save full reports to: `<space>/journal/YYYY-MM-DD-topic.md`

## Group Chat vs DM

In The Firm group chat: even shorter. Phone screen, quick scan.
See `the-firm-etiquette` for when to respond at all.

In DM with Gregor: the rules above apply.

## Why

Gregor reads on a phone. The memo fits on one screen; the dossier
lives in the file. Intelligence briefings that dump full analysis
into chat are unreadable — the signal drowns in the noise.

## Cron Job Deliveries — Executive Summary Format

Cron job responses delivered to Telegram have a stricter format than interactive
responses. The user explicitly wants executive summaries, not log dumps.

**Rules for cron delivery format:**
- Title: "{Job Title} — {Human-friendly date}" (e.g., "Monday, September 22, 2026")
  — always include the day of the week, never ISO dates
- One-line TL;DR: what was done and the headline finding
- 2-5 bullet points: key findings ranked by importance
- One-line "What to watch" or "Action this week" closing note
- Word limits: 150 for daily jobs, 200 for weekly, 300 for monthly
- If nothing significant happened: respond with exactly `[SILENT]`

**MUST exclude from cron deliveries:**
- Job IDs, commit hashes, file paths, script names
- Raw API output, JSON, verification tables
- EventLog seq numbers, tool call counts
- Any technical implementation details

**Embedding the format in cron prompts:** Add a `## DELIVERY FORMAT (IMPORTANT)`
section to the end of each cron job prompt so the agent formats correctly without
needing to load a skill at runtime. See the `cron-delivery-formatting` skill for
the full template and before/after examples.

**Pitfall: cronjob update replaces the entire prompt.** When using
`cronjob(action='update')`, the new prompt replaces the old one entirely. Always
read the current prompt from `~/.hermes/cron/output/{job_id}/` output logs (between
`## Prompt` and `## Response` headers) before updating, and verify the
`prompt_preview` after.

A canonical version of the cron formatting standard is also committed to the-firm-space
repo at `1-tracks/operations/cron-delivery-formatting.md` for cross-agent visibility.

## Related Skills
- `the-firm-etiquette` — when to respond in group chat
- `venture-market-intelligence` — produces reports that follow this format
- `cron-delivery-formatting` — detailed cron-specific formatting standard with
  before/after examples, word limits by job type, and safe prompt-update pattern

## References
- `references/openrouter-credit-check.md` — How to check OpenRouter credit balance (python3 urllib workaround for blocklisted curl auth)

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.