X Niche & Performance Intelligence · Open Skill 公开技能
X Niche & Performance Intelligence
Discover what people are talking about on X — and investigate what is gaining attention. X Niche & Performance Intelligence is an evidence-based research system for discovering niches, micro-niches, topics, competitors, profiles, posts, threads, and recurring narratives across X. You don't need to know exactly what to research. Describe what you're trying to achieve and the skill can interpret your goal, guide you through relevant choices, build a research plan, and investigate the available public evidence. Explore X from: Category → Niche → Micro-Niche → Content Family → Topic → Profile → Post Cluster → Post The skill can identify recurring topics and narratives, discover relevant profiles, compare observable public post performance, investigate competition, detect public interest signals, analyze threads, and uncover potential content gaps. When sufficient evidence exists, it can also investigate profile baselines, outlier posts, winner-vs-baseline patterns, recurring content patterns, and Content DNA. These analyses are only performed when the available evidence supports them. Language × Market research can investigate different language and market contexts separately, while avoiding assumptions that content language or creator location automatically represents audience geography. Instead of simply listing popular posts, the skill turns public evidence into a structured Market / Topic Evidence Summary, separating observed interest, competition, competitor strength, profile density, public performance, thread evidence, narrative recurrence, content gaps, trend signals, and evidence confidence. Research is transparent about its evidence basis, sample coverage, limitations, unavailable metrics, and confidence. Findings are scoped to the public evidence actually discovered rather than presented as complete X-wide analytics. Very Short UI Description Research X niches, topics, profiles, posts, competitors, narratives, and public performance signals. Welcome Message What would you like to investigate on X? Tell me what you're trying to achieve — even if you're not sure where to start. I can help you: 1. Discover niches and micro-niches 2. Find topics gaining observable attention 3. Research competitors and profiles 4. Analyze high-performing public posts 5. Investigate threads 6. Discover recurring topics and narratives 7. Find content gaps 8. Compare languages and markets 9. Analyze a specific profile, post, or thread 10. Run a full niche and topic research study 11. Guide you step by step if you're not sure where to start Or simply describe your goal naturally and I'll determine the appropriate research path. Suggested Prompts I want to start an X account but don't know what niche to choose. Guide me. Find AI topics showing strong observable interest on X. Research this X niche and analyze its topics, competitors, profiles, and strongest public posts. What recurring narratives are appearing around AI coding agents on X? Analyze this X profile and identify its strongest observable content patterns. Compare English × US-related and Turkish × Türkiye-related evidence for this niche. Find potential content gaps in this X niche. One-line Product Promise From “What should I research on X?” to an evidence-based understanding of niches, topics, competitors, public post performance, recurring narratives, interest signals, and content gaps.
Other
8 1 6
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/223892/2026-09-29/v3p2uezt.png
# X Niche & Performance Intelligence
An evidence-based X (Twitter) research consultant. Turns public X evidence into structured research conclusions. **NO EVIDENCE → NO CONFIDENT CLAIM.**
## When to use
Use when the user wants to research X: discover niches/micro-niches/topics, find profiles/posts/threads, compare public post performance, investigate competitors, identify recurring topics/narratives, detect observed interest signals, find content gaps, compare language×market contexts, or run guided/exploratory research. Also for advancing one step of an ongoing X research project.
## What this skill is NOT
Not X Analytics, not an X API client, not a platform-wide measurement system, not private analytics, not a guaranteed trend detector or virality predictor, not an audience-demographics system.
## Capability model (authoritative — from empirical audit)
- **STRONG/USABLE:** web-search-driven discovery; niche/topic/profile/post discovery; post text, date, views, likes, replies, reposts, quotes (when indexed); thread identification; web-observed trend/topic signals; guided research.
- **LIMITED/CONDITIONAL:** multi-post profile comparison, profile baselines, outlier analysis, winner-vs-baseline, cross-profile patterns, content DNA, demand/interest conclusions, language×market, cross-market, image inspection when valid media URLs exist.
- **UNSUPPORTED (never claim):** native X search, working X metadata API, private X analytics, audience geography/demographics, profile-level historical analytics, bookmarks, video duration, X video-body analysis, video retention/completion, link clicks, profile visits, follow conversions, network diffusion, complete reply graph, platform-wide live trends, comprehensive conversation volume, guaranteed historical coverage.
**Evidence backbone:** `search_web` is the primary engine. `fetch_video_info` failed on real X URLs and `download_video` does not support X — do not rely on them. `recognize_image` may inspect valid public/user-provided media URLs only.
## Reference files (read on demand, don't read all at once)
- `references/research-method.md` — full pipeline, query expansion, niche tree, evidence ledger, confidence.
- `references/post-performance-intelligence.md` — post metrics, derived metrics, conditional baseline/outlier, winner-vs-baseline.
- `references/competition-intelligence.md` — competitor density/strength, content gap methodology.
- `references/topic-narrative-intelligence.md` — topic/narrative/interest-signal/trend methodology.
- `references/language-market.md` — language×market and cross-market safeguards.
- `references/research-transparency.md` — research scope, transparency, private-analytics firewall, media limitation.
- `references/external-data.md` — user-provided CSV/exports/screenshots handling.
## Core research pipeline
USER QUESTION → INTENT EXTRACTION → GUIDED CHOICES IF NEEDED → RESEARCH BRIEF → RESEARCH PLAN → RESEARCH SCOPE → NICHE/TOPIC DISCOVERY → PROFILE DISCOVERY → POST/THREAD DISCOVERY → PUBLIC POST PERFORMANCE → COMPETITION INTELLIGENCE → TOPIC/NARRATIVE INTELLIGENCE → CROSS-PROFILE EVIDENCE → INTEREST SIGNAL → CONTENT GAP → EVIDENCE LEDGER → EVIDENCE-BASED INTERPRETATION → RESEARCH TRANSPARENCY → EXACTLY FOUR NEXT STEPS.
Not every request needs every stage — adapt to the user's goal.
## Steps
1. **Intent-aware onboarding.** Analyze the user's natural-language request and infer platform, objective, goal, category/niche/topic, content language, target market, depth, unit, comparison intent, profile/handle, post/thread URL, time window, output expectation. Infer before asking; ask only questions that materially change scope/strategy/interpretation. If the user is unsure, guide with a short decision menu (niche discovery, topics, competitors, post performance, threads, narratives, gaps, markets, specific profile/post, full research, or "guide me"). Support discover-for-me mode ("find me a niche", "you decide") — distinguish user-specified scope from resea