Skill This
AI Skill Report Card

Curating Academic Literature

A-82·Oct 28, 2026·Source: Extension-page
Quick Start13 / 15

Given a topic and raw search results, produce a structured, credibility-ranked bibliography:

  1. Filter out non-academic sources (blogs, press releases, forums, SEO content)
  2. Verify each remaining source has a traceable identifier (DOI, PMID, PMCID, ISBN)
  3. Group by stance/framework (e.g., proponents vs. skeptics vs. methodological reform)
  4. Output with full metadata and direct repository links
Recommendation▾
Add a third example showing a borderline/edge case (e.g., a source with no DOI that must be flagged unverified) to demonstrate the pitfall handling in action
Workflow14 / 15

Progress:

  • Step 1: Identify the research question and contentious sub-debates within it
  • Step 2: Triage sources into tiers (peer-reviewed journal > preprint/repository > institutional report > reputable science journalism > exclude)
  • Step 3: Extract identifiers (DOI, PMID, PMCID) for every surviving source — if none exists, flag as "unverified" or drop
  • Step 4: Summarize each paper's core claim/mechanism in one sentence, not just its title
  • Step 5: Cluster papers by argumentative position, not just chronology or source
  • Step 6: Build a comparison table contrasting frameworks/positions head-to-head
  • Step 7: Offer concrete next-step deliverables (bibliography format, deep-dive summary, specific experiment breakdown)

Triage criteria (Step 2)

  • Include: peer-reviewed journals (Springer, Oxford Academic, Taylor & Francis, Wiley), PubMed/PMC-indexed papers, institutional repositories (university thesis repos, NIH)
  • Include cautiously: preprint servers (bioRxiv, arXiv) — label as "preprint, not peer-reviewed"
  • Exclude: SEO blog posts, content farms, forum threads, unsourced popular science without citations
  • Exclude unless clearly labeled: advocacy sites with no methodology disclosure
Recommendation▾
Include a concrete output template/format (e.g., sample citation block with fields) rather than only describing what metadata to include
Examples14 / 20

Example 1: Input: List of 15 mixed search results on "plant consciousness," including PMC articles, a Psychology Today blog, and a Reddit thread. Output: 12 qualifying sources kept (all PMC/Springer/Oxford/Taylor & Francis), 3 dropped (blog, forum, uncited magazine piece). Each kept source gets PMID/PMCID/DOI, one-sentence mechanism summary, and is tagged into one of three clusters: "Pro-sentience," "Skeptical dissent," "Methodological reform."

Example 2: Input: "Find credible papers on intermittent fasting and longevity." Output: Table of RCTs and meta-analyses from journals like Cell Metabolism, NEJM, with PMIDs; excludes supplement-company-sponsored blog summaries; flags any industry-funded study as a potential conflict of interest in a footnote.

Recommendation▾
Show a 'bad output' example contrasted with a 'good output' to reinforce the credibility distinctions mentioned in Best Practices
Best Practices
  • Always surface the identifier (DOI/PMID/PMCID) — a citation without one is not yet verified
  • State the core mechanism/claim in plain language, not just the paper title
  • Explicitly separate "what proponents argue" from "what critics argue" — don't flatten debates into a single consensus
  • Note publication venue prestige/peer-review status explicitly (e.g., "Springer Nature, 2024" vs. "self-published PDF")
  • When a claim lacks empirical confirmation, say so directly rather than implying consensus
  • End with actionable next steps (bibliography formatting, targeted summaries, deeper dives) rather than a flat list
Common Pitfalls
  • Don't count citation count or topical relevance alone as "credibility" — check peer-review status
  • Don't merge dissenting and supporting papers into one undifferentiated list — the tension between positions is the valuable signal
  • Don't fabricate or guess at DOIs/PMIDs — if unverifiable, label clearly as "identifier not found"
  • Don't let promotional/advocacy framing from source abstracts leak into neutral summaries
  • Don't treat preprints or non-peer-reviewed working papers as equivalent to published, peer-reviewed findings
0
Grade A-AI Skill Framework
Scorecard
Criteria Breakdown
Quick Start
13/15
Workflow
14/15
Examples
14/20
Completeness
18/20
Format
15/15
Conciseness
13/15
Save to My Skills