X/Twitter Research with Cited Post Evidence
Research X topics, threads, and accounts with cited posts, deliberate sampling, and clear limits on sentiment claims.
Before & after
Same prompt, same agent — the only difference is the skill.
I can help you research what people are talking about regarding AI safety. You'd want to search for relevant discussions and see what's trending. Do you want recent posts only or a broader historical look? Are you interested in expert takes or grassroots sentiment? There are a lot of tweets about AI safety, so if you could narrow down the specific angle you're most interested in, I could help synthesize the main themes.
X RESEARCH REPORT: AI Safety Sentiment
→ DATA COLLECTION
Time window: Last 7 days (Aug 24-31)
Posts analyzed: 487 relevant tweets (filtered by AI + safety keywords)
Engagement total: 14.2M impressions, 820K interactions
Top voices: Mix of researchers (45%), policymakers (20%), builders (18%), others (17%)
→ SENTIMENT BREAKDOWN
Positive (we can solve this): 31% of posts • cautiously optimistic tone
Negative (doom-focused): 38% of posts • warnings about AGI risks
Neutral (factual discussion): 21% of posts • debate, no stance
Sarcastic/critique of discourse: 10% of posts • meta-commentary
→ TOP 3 THEMES (by post volume & engagement)
1. "Alignment research bottleneck" (142 posts, 2.3M impressions)
Common claim: "Not enough researchers focused on interpretability"
Refrain: Key labs need 10x hiring in safety teams
Notable quote: "Lab A claims progress but methodology is closed-box"
2. "Regulation timing urgency" (118 posts, 1.8M impressions)
Common claim: "EU AI Act lags behind capability acceleration"
Refrain: Governance frameworks outdated within 6 months
Notable quote: "By the time rules pass, the tech is obsolete"
3. "Corporate accountability gap" (97 posts, 1.6M impressions)
Common claim: "Self-regulation doesn't work; independent audits needed"
Refrain: Open-source models bypass safety constraints
Notable quote: "Accountability only happens after incident, not before"
→ INFLUENCER HIGHLIGHTS
Researcher A (15k followers, high resonance): Posted thread on interpretability costs
Policy Org B (84k followers): Analysis of pending legislation vs. labs' capabilities
Builder C (3.2k followers, engaged community): Practical safety guardrails for production systems
→ ACTION INSIGHT
Conversation gaps: Underrepresented – concrete safety solutions (mostly problem-focused)
Opportunity: Your content on working safety implementations would stand out (low noise)About this skill
“What are people actually saying?” is hard to answer from a handful of viral posts or an incomplete reply tree. This skill turns accessible X/Twitter posts into a cited research brief, showing the recurring arguments, useful audience language, and limits of the evidence.
What it does
- Researches a topic, a specific account, or an announcement thread through an available, authorized terminal retrieval route.
- Builds searches around names, alternative terms, objections, and date windows so one phrasing does not dominate the findings.
- Separates original posts, replies, reposts, and quote posts, then removes duplicate evidence.
- Reconstructs accessible thread context and distinguishes direct replies from nested discussion.
- Reports sentiment with explicit counts and denominators, without treating engagement-ranked results as public opinion.
- Attaches post links, timestamps, and observed engagement context to material findings.
What you get
A concise research brief with cited findings, audience wording, counterarguments, sample counts, and access limitations. For deeper work, you also get a normalized post ledger and a record of the queries and collection boundaries. Retrieval requires an accessible public source or an existing authorized integration; no paid API bypass or complete thread access is promised.
Best for
Marketers studying customer objections, builders investigating product feedback, and researchers examining public account activity or reactions to a launch.
SKILL.md preview
---
name: x-twitter-research-with-cited-post-evidence
description: Research X/Twitter topics, threads, and account activity through available authorized terminal access, then deduplicate posts, reconstruct discussion context, and produce a cited synthesis with explicit sampling limits. Use when the user says "research this topic on X", "what are people saying about this launch?", "analyze this thread and its replies", "check this account's recent posts", "find repeated customer complaints", "my X searches keep showing the same viral posts", or "is this backlash widespread or just one thread?".
---
# X/Twitter Research with Cited Post Evidence
… (sign up to view the full skill)More research & data skills
View all Research & Data skills →Source-Reconciled Research Brief
Turn supplied links into a concise brief with claim-level citations, reconciled source conflicts, and explicit evidence gaps.
Source-Traced Research Bundle
Compile many sources into one bundle where every claim is atomic, sourced, dated, tiered, and conflict-checked, in Markdown or JSON.
A/B Test Power and Validity Designer
Exact A/B sample sizes and MDE with math shown, plus the traps: unit mismatch, multiple variants, peeking, guardrail power, and SRM.
Research Routing and Cited Briefings
Route research to the right method, then produce a cited briefing with evidence gaps, disagreements, and a source audit.