---
name: "Force x_search + web_search in parallel, never default-web"
description: "Use when a research brief must force x_search + web_search in parallel and never default-web."
---

# Force x_search + web_search in parallel, never default-web

## When
Use when live X firehose and indexed web both matter (competitive, mentions, pricing).

## Stop
Do not default to web_search. Do not post, buy, or email from this skill. If a source is missing, say so — no filler.

## Steps
1. Put this in the skill instructions: `Use x_search for live posts, sentiment, employee chatter, customer complaints. Use web_search for news, docs, pricing pages, job listings. Do both in parallel. Do not default to web_search.`
2. Parallel gather (example competitive brief): X official + employee accounts; web news/press; web job postings; X customer complaints; web pricing page.
3. Mark every claim HIGH/MEDIUM/LOW and cite the URL or post. If a source is missing, say so — no filler.
4. Output a scannable brief, not a narrative: Situation, Signal, So what, Recommended action, Watch list. Cap length (e.g. 400 words).
5. Recency: distinguish 2024 vs 2026 findings. Flag a narrative that appears in the last 2 hours but not the 24-hour baseline.
6. Save as `/brief` or `/compete`. Invoke with `/`. Do not post, buy, or email from this skill.

## Load
x_search + web_search in parallel. /brief or /compete. Cite URLs. Cap ~400 words.