Source-Reconciled Research Brief
Turn supplied links into a concise brief with claim-level citations, reconciled source conflicts, and explicit evidence gaps.
Before & after
Same prompt, same agent — the only difference is the skill.
I can help with that. To do it properly, I'd need to know a bit more. Are you interested in US regulation specifically or global? Which countries matter most to you? And what's the use case — are you writing for a policy audience, your board, or a general audience? Also, do you have specific sources you want me to include, or should I find them? Once I have those details, I can put together a summary.
BRIEF: AI REGULATION LANDSCAPE (2026) SUMMARY: Three key shifts emerged in first half 2026: US enforcement under Executive Order 2025-002 intensified, EU AI Act Phase 2 penalties began, and six countries rolled out binding liability frameworks. No single framework dominates; firms now operate under parallel rules. KEY POINTS: • US: FTC actions on model training data + transparency claims (Source A) • EU: 4 billion euro fine precedent set March (Source B) — confirmed escalating after pilot phase • UK, Canada, Australia: Adopted liability model (who pays when AI fails?) by June (Source C) — diverges from EU • China: Announced data residency requirement for large models (Source D) — impacts global deployment CONTESTED: Boundary between US state rules (CA stricter than FL) versus federal preemption — Congress has draft but no vote scheduled (Source B says Q4 2026, Source C says 'unclear') OPEN QUESTIONS: • Will international standard emerge by 2027 or splinter further? • Will liability framework apply retroactively to models trained 2023-2024? • Which definition of 'high-risk AI' sticks? (Still varies by region) OWNER: Researcher A USE: Cite this brief in your Dec compliance memo; revisit in 60 days
About this skill
Five links can leave you with five incompatible versions of the story. This skill turns your supplied sources into a brief that shows what the evidence supports, which disagreements actually matter, and what you still need to verify before acting.
What it does
- Attaches citations to individual factual claims, with page, section, or timestamp locators when available.
- Distinguishes independent corroboration from articles repeating the same announcement or dataset.
- Checks dates, definitions, populations, and measurement methods before calling two claims contradictory.
- Flags consequential claims supported by only one source and separates reported facts from interpretation.
- Keeps inaccessible links and unanswered questions visible without filling gaps from memory.
- Compresses the findings around your meeting, question, or decision instead of summarizing each article in turn.
What you get
A concise briefing with a bottom-line finding, supported key points, material disagreements, a source register, and targeted verification questions. Quick claim checks receive a compact verdict rather than a full report.
Best for
Analysts, operators, founders, and meeting owners who already have several sources and need a defensible synthesis before a decision, presentation, or discussion.
SKILL.md preview
---
name: source-reconciled-research-brief
description: Compiles supplied links or excerpts into a concise research brief by tracing claims to original evidence, checking comparability, and reconciling or preserving material disagreements. Use when the user says "turn these links into a brief", "these articles contradict each other", "what do these sources actually support?", "I need a sourced summary before my meeting", "my research notes keep turning into a wall of text", "check this claim against these sources", or "what is still unknown?".
---
# Source-Reconciled Research Brief
… (sign up to view the full skill)Featured in
· 7 min read
15 Claude Skills Examples: Real SKILL.md Files by Use Case (2026)
· 6 min read
How to Install Agent Skills in Claude, ChatGPT, Cursor, Codex & Gemini (2026)
· 7 min read
15 Best Claude Skills to Install in 2026 (for Writing, Sales & Productivity)
· 8 min read
What Are Agent Skills? The Complete Guide to SKILL.md (2026)
More research & data skills
View all Research & Data skills →Source-Traced Research Bundle
Compile many sources into one bundle where every claim is atomic, sourced, dated, tiered, and conflict-checked, in Markdown or JSON.
A/B Test Power and Validity Designer
Exact A/B sample sizes and MDE with math shown, plus the traps: unit mismatch, multiple variants, peeking, guardrail power, and SRM.
Research Routing and Cited Briefings
Route research to the right method, then produce a cited briefing with evidence gaps, disagreements, and a source audit.
X/Twitter Research with Cited Post Evidence
Research X topics, threads, and accounts with cited posts, deliberate sampling, and clear limits on sentiment claims.