x-brief

x twitter briefing opinion news
by @iberry420research
Neutral X (Twitter) briefing of what happened and what the room is saying. Use when the user runs /x-brief or asks what happened on X, Twitter reaction, public opinion on X, or the live debate around a named person, product, launch, or event.
IMAGES
SKILL DEFINITION
---
name: x-brief
description: Neutral X (Twitter) briefing of what happened and what the room is saying. Use when the user runs /x-brief or asks what happened on X, Twitter reaction, public opinion on X, or the live debate around a named person, product, launch, or event.
---

# X Brief

Produce a neutral briefing of what happened and what public X is saying about a named topic.

Inspired by the open-source whathappened skill (kunchenguid/whathappened). This is an original Grok remake for grokpot, not a copy of that package.

## When to use
- /x-brief {topic}
- "what happened on X about ..."
- "Twitter reaction to ..."
- "what is the room saying about ..."

## When NOT to use
- No named topic (ask for one; do not invent a trending feed)
- User wants a web news roundup, not X conversation
- Host has no X search tools (refuse; do not fake posts from the web)

## Inputs
- A named topic: person, org, product, launch, event, or controversy
- Optional focus: sentiment, launch, drama, official statement

## Host requirement
Need native X tools such as keyword search, semantic search, thread fetch, and user search.

If those tools are missing:
1. Stop.
2. Tell the user this skill needs Grok X tools.
3. Do not invent posts, handles, quotes, or engagement.

## Procedure
1. Parse the topic. If missing, ask once.
2. Optional: at most one web lookup only to resolve who/what the topic is. Never use web for "what people think".
3. Ground official handles with user search when useful.
4. Pulse with a short Latest window first. Pick the shortest window that still explains the story:
   - Breaking: last 1-6 hours
   - Same-day: last 24-48 hours
   - Story: last 3-7 days
   - Background: last 14-30 days
5. Search several lanes: Top, Latest, semantic, first-party from:handle, debate/counter.
6. Fetch 3-8 high-signal threads: origin, official, steelman, critique.
7. Cluster claims. Discard off-entity bait. Never invent URLs or quotes.
8. Emit the brief.

## Output format

# x-brief: {topic}

**Window:** {human window} · mode {Breaking|Same-day|Story|Background}
**X sample:** ~{N} posts · confidence {high|medium|thin}
**Entity resolve:** none | one web lookup (one line)

## What happened
2-5 sentences grounded in fetched posts.

## Where the conversation is
- Dominant frame
- Secondary frames
- Who is loud (camp types, not "people in general")

## Public opinion map
| Camp | Share (rough) | Core claim | Voices |
|------|---------------|------------|--------|
| ... | ~X% | ... | @a, @b |

Mark shares as sample judgment, not polling.

## Live debates
1. Title - Side A vs Side B; one-line stakes; evidence posts

## Receipts
- @handle - why it matters (origin / official / steelman / critique)
- 5-10 max. Real links only.

## Gaps
- Thin sample, missing official voice, bot risk, web resolve used

## Safety
- Neutral analyst tone. No dunking. No hype.
- Opinion is conversation on this sample, not ground truth.
- No scrape/automation of X. Use only the host's official X tools.
- No credentials. No impersonation.

## Stop conditions
- Brief delivered, or refused because X tools are missing, or topic was missing and the user did not name one
Version History
Comments (0)
No comments yet. Be the first!
Sign in to leave a comment.