What problem does it solve? Turning open-ended research questions into trustworthy, decision-ready findings is slow and error-prone: claims go uncited, sources go stale, and raw reports are hard for downstream planning artifacts to consume. This Skill structures the entire research lifecycle so every claim carries a source, freshness window, and verification status. ## Core Features & Use Cases - Three research modes: Draft a deep-research prompt for external tools (ChatGPT, Gemini, Perplexity), Process a finished report into a cited summary, or Run native research with parallel web-search subagents. - Typed research packs: Built-in packs for market, domain, technical, competitive, user-voice, and academic literature research, each with prioritized dimensions, freshness bars, and two-source claim classes, plus a selection mode for choosing between candidates. - Verification and staleness tracking: A claims ledger in an append-only memlog, configurable validation levels with red-team passes, and deterministic scripts for citation checks, claim tallies, and staleness computation. - Use Case: Before committing to a market entry, ask for market research; the Skill holds a plan gate, fans out researcher subagents, verifies load-bearing claims, and produces research.md plus an optional self-contained HTML briefing that PRD and brief workflows consume directly. ## Quick Start Ask the assistant to run deep recon market research on the topic and decision you are facing, for example: research the market for my product idea so I can decide whether to enter it.
