AI research & analysis skills for agents
The failure mode of AI research is confident nonsense. These skills force the disciplines that prevent it: claims separated from framing, sources ranked by incentive and methodology, falsifiability stated up front.
Benchmarked on suites where outputs are verifiable — extraction, source triage, fact-check planning.
16 skills · 1 benchmarked with Lift · sorted by measured performance, then Trust.
Scores a scoped set of competitors across nine weighted dimensions using evidence-anchored 1–5 rubrics and a strategic tension plot to produ
Analyzes and categorizes competitors to define a focused benchmarking scope based on positioning, niche, and market data.
Assembles scored competitor profile cards into a decision-grade competitive report with landscape maps, benchmarking matrices, white-space a
Runs multi-source web research using firecrawl and exa tools, then synthesizes findings into a cited report with source attribution.
Diagnose and propose a no-code product-rule plan to resolve stalled, looping, or over-recovered issue trees
AI-powered neural search tool for web content, code examples, company research, and professional profiles using Exa MCP
Analyzes YouTube video data to identify top-performing content and strategies within a niche.
Improve Archify diagram accuracy, visual quality, and authoring speed through systematic experimentation and evaluation
Extracts business domain knowledge from a codebase and generates an interactive flow graph to visualize domains, business flows, and process
This skill helps users scrutinize AI-generated content before acting on it by appending follow-up questions to verify facts, probe reasoning
Guides how to search X/Twitter posts via the x_search tool with date bounds, trusted account scoping, operator-based queries, and reconcilia
Builds a discounted cash flow model to estimate a public company's intrinsic value per share, with documented assumptions, sensitivity analy
Hyperplan is an adversarial multi-agent planning skill that orchestrates a team of hostile agents to ruthlessly critique and debate a planni
Generates a narrative 'Journey Into [Project]' report analyzing a project's full development history from claude-mem's timeline, including a
This skill activates a planning consultant that grounds in the codebase and creates a decision-complete work plan for a worker to execute wi
Build and query AI-powered knowledge bases from observation history to create custom brains and ask questions about past work patterns
FAQ
What is an AI summarizer skill?
A compression contract: keep numbers and obligations, drop hedging, label uncertainty — your agent applies it to any document or thread.
Can an agent do deep research with these?
The research skills contribute method — query decomposition, source ranking, synthesis with citations — on top of whatever browsing tools your agent has.