research-lookup

Trigger phrases

Phrases that activate this skill when typed to Claude Code:

• look up research on
• find current information about
• search academic papers on
• deep research on
• verify this scientific claim

What it does

research-lookup is a Claude Code skill from K-Dense AI’s scientific-agent-skills repo. It turns Claude into a research router that automatically selects the fastest, most appropriate backend for any research query: parallel-cli for quick web searches, the Parallel Chat API for deep multi-source research reports, or Perplexity sonar-pro-search for academic paper queries.

A session produces sourced answers — not memory-based responses. The routing logic means you don’t have to decide which backend to use; the skill picks based on query type and available credentials.

When to use it

Reach for it when:

• You want to verify a scientific claim with current sources rather than relying on Claude’s training data
• You need a quick factual answer grounded in live web or academic sources
• You’re inside another skill (e.g., scientific-writing or literature-review) and need a fast sub-query for current data

When not to reach for it:

• Systematic literature reviews with documented search methodology — use literature-review
• Direct database queries with specific DOI/PMID lookups — use paper-lookup

Install

Copy the SKILL.md from K-Dense AI’s research-lookup folder into .claude/skills/research-lookup/ in your project. Requires parallel-cli for the primary backend; PARALLEL_API_KEY and OPENROUTER_API_KEY are optional for deep research and academic modes.

Trigger phrases: “look up research on”, “find current information about”, “search academic papers on”, “deep research on”.

What a session looks like

A typical session has three phases:

1. Query classification. Claude determines whether the query is a quick factual lookup, a deep research question, or an academic literature search, and selects the appropriate backend.
2. Retrieval. The selected backend runs the search — fast web for quick lookups, multi-source deep research for comprehensive questions, Perplexity academic for paper searches.
3. Sourced output. Results are returned with source links and publication context. For deep research mode, a structured report with citations is generated.

Receipts

Where it works well:

• Fact-checking statistical claims in manuscripts — quick web mode returns current figures with sources
• Sub-queries from within other skills, where a targeted lookup is needed mid-workflow without switching context

Where it backfires:

• Without parallel-cli installed, the primary backend is unavailable and the skill falls back to slower alternatives
• Deep research mode can be verbose — useful for reports but overkill for simple factual queries

Pattern that works: this skill is most valuable as a component called by other skills (scientific-writing, market-research-reports) rather than standalone; its auto-routing saves context-switching when the right backend isn’t obvious.

Source and attribution

Originally authored by K-Dense Inc.. The canonical SKILL.md lives in the research-lookup folder of their public scientific-agent-skills repository.

License: MIT. Install, adapt, and redistribute with attribution preserved.

This page documents the skill from a practitioner’s perspective. For the formal spec and any updates, defer to the source repo.