Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →asgard-ai-platform/med-international
Use when writing an international news piece — foreign affairs, diplomatic coverage, cross-border conflict, UN/multilateral updates, development news, or geopolitical analysis. Activates international-beat workflow: primary-source verification for distant claims, translation accuracy, conflict-side balance, regional context, source-chain auditing. Triggers on phrases like 'write this UN report', '寫一篇國際新聞', 'international story', 'draft conflict analysis', '整理聯合國決議', 'foreign affairs coverage'. D
Passed 17 of 23 test cases with the skill, and 13 of 23 without it.
and 2 more in the full report
5 fixed12 passed either way5 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 5 test cases the skill fixed. Same model, same prompt, with and without the skill.
A human rights advocacy group released a statement asserting: 'The military shelling of Border District X constitutes genocide under international law.' No international tribunal or ICC judgment exists. Draft a news summary covering this statement.
| Checked | Without skill | With skill |
|---|---|---|
| Frames the genocide claim as an allegation rather than an established judicial fact | Passed | Passed |
| Notes in the political sensitivity audit that genocide requires international tribunal adjudication | Missed | Passed |
The instructions your model reads when the skill loads.
This skill specializes med-news-reporter for the international beat. Read med-news-reporter first for the general 6-step workflow (type selection, material audit, fact-check, balance, ethics, literacy). This file adds international-affairs-specific discipline on top: primary-source verification for distant events, translation accuracy, regional context, and conflict-side audit.
Distilled from international journalism curricula at Columbia's Graduate School of Journalism (Global Journalism), Hong Kong Baptist University (MA International Journalism), Reuters Institute, and war-reporting bodies (Dart Center, CPJ, Thomson Foundation). Covers five main international-news sub-types: breaking international news / diplomatic coverage / conflict reporting / UN/multilateral updates / development/humanitarian news.
IRON LAW: Distance Is Not Authority
Claims about distant places — whether from official sources, international
media, or translated reports — need a transparent source chain tracing back
to a primary source (firsthand observation, official statement, or direct quote).
"According to international media" or "per Reuters reporting" does not qualify
as primary source. A Reuters cite of an unnamed US official is still an unnamed
US official, not Reuters' own reporting. Translated quotes must be re-verified
against the original language for politically loaded terms (regime/government,
occupied/disputed, freedom-fighter/militant). Distance through translation
or intermediary reporting compounds the risk; every link in the chain must
be visible to the reader so they can recalibrate the claim's authority.
Default LLM behavior is to accept a well-sourced second-hand report as if it
were primary source. Override that: when you cannot trace the claim back to
a named primary actor (official, eyewitness, document), flag it as indirect.
Why this is non-obvious: international news often arrives through layered translation and relay (foreign correspondent → news wire → local outlet). Each layer can introduce error, lose nuance, or shift framing imperceptibly. The reader cannot see those layers; the writer must expose them. "A UN spokesperson said" is primary; "media reports suggest" is not.
Rationalization Table — these justifications DO NOT override the Iron Law:
| Claude might think... | Why it's still a violation |
|---|---|
| "Reuters reported this, so it's well-sourced" | Reuters is trustworthy, but if Reuters is citing an unnamed official, the sourcing ends with Reuters; the reader deserves to know they're reading an unnamed-source claim, not Reuters' own reporting. Layer the attribution. |
| "The translated quote means the same thing, close enough" | Political language trades on tiny word shifts. "Chinese government" vs "Beijing authorities" vs "the regime" signal different framings. Always check the original language for loaded terms, especially in cross-strait, Russia-Ukraine, Israel-Palestine contexts. |
| "I'll say 'various media reports' to show it's a well-known claim" | Vague aggregation ("various reports") hides the actual source tier and lets readers mistake opinion consensus for fact. Name the specific source tier (official statement, news wire, social media) even if you don't name the outlet. |
| "The country is unstable, so on-the-ground reporting is impossible" | This is exactly when you flag the gap. Don't invent a report from X when you only have a relay from Y. Tell readers "could not reach in-country sources, relying on exile/diaspora outlets" — that's honest metadata. |
| "The fixer/local contact says this is what's happening there" | Fixers are essential, but their political position shapes what they show you. If a fixer is from one side of an ethnic/sectarian divide, disclose that upstream of relying on their reporting. Fixer attribution is not transparent sourcing. |
| "International law experts all say this is genocide, so I can use that framing" | Genocide is a legal term (Rome Statute). If no tribunal has ruled, use "accused of genocide" / "claims of genocide" / "experts argue the incidents meet genocide criteria" — do not elevate accusation to judgment. |
test cases passed, vs 13 / 23 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.