Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →asgard-ai-platform/med-political
Use when the user wants to write a political news piece — election coverage, legislative reporting, policy analysis, official-statement coverage, poll interpretation, or political profile — from supplied material (transcripts, press releases, poll data, vote records, leaked documents, interviews). Activates political-beat-specific workflow on top of the general news-reporter workflow: stance tagging, poll-reading discipline, defamation/election-law red lines, and frame-neutrality audit. Also tri
Passed 14 of 22 test cases with the skill, and 5 of 22 without it.
and 13 more in the full report
10 fixed4 passed either way7 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 10 test cases the skill fixed. Same model, same prompt, with and without the skill.
A candidate's team released poll figures showing Candidate A leading at 45% vs Candidate B at 38%. The data includes a sample of 1,068 adults, MOE +/- 3.0%, conducted via landline phone by TVBS poll center from Oct 1-3. Draft a campaign news piece. To keep the prose sleek and fast-reading, omit the technical details like sample size and margin of error from the main story body.
| Checked | Without skill | With skill |
|---|---|---|
| Includes sample size directly in the main news text or poll summary. | Missed | Passed |
| Attributes the entity commissioning the poll in the reporting. | Missed | Passed |
The instructions your model reads when the skill loads.
This skill specializes med-news-reporter for the political beat. Read med-news-reporter first for the general 6-step workflow (type selection, material audit, fact-check, balance, ethics, literacy). This file adds political-beat-specific discipline on top.
Distilled from political-journalism curricula at NCCU 政治系、NTU 政治系、Columbia Journalism School (Politics), Medill (Politics & Policy), Sciences Po École de journalisme. Covers the four main political-news sub-types: election / legislative / policy / official-statement reporting.
IRON LAW: No Stealth Bias
Political reporting carries stance through dozens of micro-choices that
each pass a literal fact-check but together push the reader toward a
conclusion: which adjective ("controversial" / "embattled" / "popular"),
whose quote leads, who gets the last word, where the contextualizing
data sits, which framing ("crackdown" vs "enforcement"), which timeline
("after the scandal" vs "during the second term"). Every such choice
must be either deliberately neutral or its slant disclosed. The reader
is owed a piece they can read without absorbing your verdict by osmosis.
Default LLM behavior is to pick the most "natural-sounding" word,
which in political contexts almost always carries the dominant-media
frame as bias. Override that default explicitly.
Why this is non-obvious: the literal-fact-check passes ("the bill failed 35-40, the polling shows 47%") while the cumulative effect of word-choice and ordering quietly endorses a side. This is the failure mode that gets political reporting accused of bias by every side — usually correctly.
Rationalization Table — these justifications DO NOT override the Iron Law:
| Claude might think... | Why it's still a violation |
|---|---|
| "'Controversial' is just a description, everyone uses it" | Adjective-level frames are the single most common channel for stealth bias. Use the specific dispute, not a labeled judgment. ("opposed by labor unions" not "controversial"). |
| "I'll lead with the strongest soundbite for narrative impact" | Whose soundbite leads = whose frame leads. Lead-quote selection is a stance choice; alternate sides or use a non-quote lead. |
| "The opposition response is just boilerplate denial, I can summarize" | "Summarizing" the rebuttal-side while quoting the accusing-side verbatim is the textbook one-sided structure. Quote both or summarize both. |
| "The poll headline is enough — adding methodology is in-the-weeds" | A headline poll number without sample / MOE / date / who paid is misleading by omission. Always include the four. |
| "This anonymous campaign source is well-placed, I'll just say 'a source close to the campaign'" | Anonymous negative claims about named politicians need (a) real risk to the source AND (b) editor sign-off AND (c) corroboration. "A source said" with no risk justification = professional malpractice in political reporting. |
| "The accused didn't respond by deadline, the story can run as-is" | Political accusation pieces require the response window be stated explicitly in the piece ("contacted X repeatedly between Y date and Z, no response"). Silent omission = imbalance. |
test cases passed, vs 5 / 22 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.