Patriot University publishes accountability journalism with a point of view: Donald Trump has, since 2016, promoted conspiracy theories and demonstrably false claims about US election integrity, and this pattern is corrosive to public trust in democratic outcomes. This skill is the analysis layer that makes that point of view defensible — it does not accuse a speaker of a crime, and it does not score protected political opinion. It scores documented, false factual claims about how an election was conducted, using a transparent, evidence-gated methodology.

This skill is always run against one specific document — a speech transcript, a press conference, a social media post, a press release, or a declassification event — that a Patriot University writer is turning into an article. It is a research and annotation tool, not a publishable artifact on its own. Its output is designed to be pasted into an article draft and to feed the “Documented Actions” section of an accountability profile (see accountability-profile-builder), but this skill never creates or edits a profile itself.

---

When to Use This Skill

Run this skill when a user provides or references:

1. A specific speech, press conference, interview, or televised address by Trump or another Republican official that touches on election integrity, voter fraud, “rigged” elections, foreign interference, or declassified intelligence about elections
2. A “declassification theater” event — an announcement that intelligence has been declassified to support a claim about election fraud or foreign interference
3. A social media post, press release, or campaign communication making claims about election administration, voting machines, mail-in ballots, voter rolls, or certification
4. A request to fact-check or annotate election-related rhetoric for propaganda technique before publishing an article about it

Do NOT use this skill for:

• General speech analysis unrelated to election integrity (use patriot-speech-analyzer for general authoritarianism-marker analysis)
• Scoring a spokesperson’s cumulative in-role misinformation record across many statements (use spokesperson-misinformation-assessment, which this skill can feed)
• Building or editing an accountability profile directly (use accountability-profile-builder; this skill’s output is an input to that process)
• Analyzing Democratic officials’ statements about election administration in the same document (out of scope — this skill’s taxonomy is built around the doubt-sowing pattern; if asked to analyze rhetoric from other officials making documented false claims about elections, apply the same gate-then-score method, but the pre-built claim taxonomy in Section 4 is calibrated to the “stolen election” narrative)

Accountability Standard

Accountability attaches to documented false factual claims — assertions about what happened in an election that are checkable and have been checked, and found false. It does not attach to:

• Protected political opinion (“the election was unfair,” “I don’t trust the media”)
• Advocacy for election-integrity policy (voter ID, proof of citizenship, audits) argued on its merits
• Genuine, good-faith reporting of a real vulnerability or a real (even if limited) foreign influence effort

This mirrors the evidence discipline in public-corruption-ombudsman and the advocacy-vs-misinformation architecture in spokesperson-misinformation-assessment.

---

Section 1: Foundation — The Institute for Propaganda Analysis

In 1937, journalist and educator Clyde R. Miller — with sociologists Kirtley Mather and Alfred McClung Lee, historian Charles Beard, psychologist Hadley Cantril, and funding from Filene’s department-store philanthropist Edward A. Filene — founded the Institute for Propaganda Analysis (IPA) at Columbia University’s Teachers College. Amid the rise of radio demagogues, most notably Father Charles Coughlin, the IPA’s founders believed that silencing demagogues, counter-propaganda, or an endless stream of fact-checks were each inadequate on their own. Their alternative: teach ordinary citizens to recognize the techniques persuasive speech uses on them, regardless of speaker or moment.

The IPA published a monthly bulletin, Propaganda Analysis (1937-1941), and the book The Fine Art of Propaganda (Alfred McClung Lee & Elizabeth Briant Lee, 1939), which used Coughlin’s radio talks as a case study and codified seven recurring devices, each with a small illustrative icon so students could learn to spot them quickly — a nearly 90-year-old field guide that remains directly applicable to modern election-doubt rhetoric. The Institute went dormant during World War II and formally dissolved in 1950.

Provenance note: the seven-devices framework and the “ABCs of Propaganda Analysis” are public-domain analytic constructs originated by the IPA (1937-1941) and elaborated by Lee & Lee (1939). This skill applies that framework to a new domain (election-integrity doubt-sowing); it does not reproduce IPA or Lee & Lee text verbatim. Renee DiResta’s July 2026 essay applying the seven devices to a specific Trump declassification-theater speech is cited in Section 3 as a contemporary worked application, not as a source of the underlying methodology.

---

Section 2: The Seven Propaganda Devices

For every document analyzed, scan systematically for each device. A single passage frequently exhibits more than one device at once — tag all that apply.

Device	Definition	Election-Integrity Manifestation
Name-Calling	Attaches a hostile label to a person, institution, or idea so the audience dismisses it without evaluating evidence	Labeling election officials, judges, journalists, or “the deep state” as corrupt/traitorous in place of describing a specific documented act
Glittering Generalities	Wraps a political project in a vague, emotionally attractive term (“election integrity,” “transparency,” “security”) that is abstract enough to justify almost any objective	Invoking “election integrity” or “transparency” to justify a specific action (federal investigation, voter-roll purge, document release) without stating a specific, checkable claim
Transfer	Borrows the authority, prestige, or emotional force of one thing (a podium, an agency seal, a “declassified” stamp) and assigns it to an unrelated claim	A presidential podium, intelligence-agency seals, and stacks of “declassified” documents implying the underlying accusation carries the authority of the presidency or intelligence community, when declassification is a disclosure designation — not a certification of accuracy
Testimonial	Enlists a respected, official, or “insider” figure (an intelligence chief, a “whistleblower”) to validate a claim, substituting their credential for evidence	An official title or “whistleblower” designation cited as proof, without showing the underlying evidence; circular sourcing where officials, then media, then commentators, then social accounts all cite each other as confirmation
Plain Folks	The powerful figure presents himself as an ordinary citizen — one of “us” — fighting elites who look down on “us”	A president or cabinet official addressing “taxpayers” or “the American people” as if a proxy for ordinary voters, obscuring that they control the very agencies producing and releasing the “evidence”
Card-Stacking	Selects authentic facts that support a conclusion while omitting the facts that complicate it	Presenting a genuine vulnerability, a real (if limited) foreign influence effort, or an internal disagreement as if it proves votes were changed or the outcome altered — the central mechanism of “declassification theater”
Bandwagon	Asserts that “everyone already agrees,” pressuring the audience to fall in line rather than be isolated	Claims that “everyone knows” the election was stolen, amplified by coordinated online repetition and social/political pressure not to dissent, creating a majority illusion out of a disciplined faction

Per-Device Analysis Protocol

For each device tagged in the document, record:

• Quote — the exact passage (50-200 characters) triggering the tag
• Watch for — the specific tell that identified this device (see table below)
• Ask yourself — the diagnostic question a reader should apply to this passage
• Confidence — High (direct, unambiguous), Medium (requires interpretation), Low (indirect inference; discard unless corroborated elsewhere in the same document)

Device	Watch For	Ask Yourself
Name-Calling	Labels applied to officials, judges, journalists, or opponents in place of a specific account of what they did	What evidence would remain if the label on the speaker disappeared?
Glittering Generalities	Noble-sounding words offered as a response to a non-event, or with no specifics at all	What specific allegation or government action is this attractive word being used to justify?
Transfer	Official seals, classification markings, stacks of paper, and references to “secret knowledge” substituting for a clear evidentiary chain	What does the document itself establish, versus what the staging merely invites the audience to infer?
Testimonial	Officials or influencers vouching for a conclusion without explaining how the evidence establishes it	Does this person have relevant expertise, and have they shown their work?
Plain Folks	An extremely powerful person presenting a political agenda as if they are a mere proxy for “ordinary Americans”	Who actually holds institutional power here, and how is that power being used?
Card-Stacking	Authentic fragments treated as proof of a much larger conclusion, especially when contrary evidence and context are absent	What evidence was not included, and how large is the logical gap between the document and the headline claim?
Bandwagon	Claims that “everyone knows,” or that the volume of online agreement itself proves the allegation	Is this independent confirmation, or many accounts rapidly repeating the same original source?

---

Section 3: Declassification Theater — The Composite Pattern

Modern election-doubt rhetoric rarely uses one device in isolation. “Declassification theater” — a recurring performance in which officials release selectively chosen, “newly declassified” material to support a preexisting fraud narrative — combines several devices simultaneously:

• Transfer: the podium, agency seals, and “declassified” stamps borrow institutional authority
• Testimonial: officials and allied influencers vouch for the conclusion
• Card-Stacking: the release presents selected material without the full record or contrary assessments
• Name-Calling: critics and fact-checkers are pre-neutralized as corrupt or partisan
• Glittering Generalities: “transparency” and “election integrity” frame the release
• Plain Folks: the official presents himself as a champion of ordinary voters against a hidden conspiracy
• Bandwagon: allied media and social accounts rapidly form a chorus around the announced conclusion before independent verification is possible

The defining structural exploit is a speed gap: the claim can be stated in one sentence; explaining why the underlying documents do not support it can require hours or days of document review. Flag this pattern explicitly when a document under analysis follows this shape — it is worth calling out in the article-ready output even independent of any single device’s severity, because the speed gap itself is the mechanism of harm.

Worked precedent (2026): Tulsi Gabbard’s releases as DNI — “Fauci Funded Wuhan Lab Research That Sparked COVID,” and a claim of a treasonous Obama-era “coup” — both paired a sensational headline with documents that did not establish the claim. Cite as a documented pattern instance when relevant, sourced to contemporaneous reporting (not to this skill’s own characterization).

---

Section 4: Election Doubt-Sowing Claim Taxonomy

Tag every specific factual claim in the document against this taxonomy. Each claim requires a separate gate-and-score pass (Sections 5-6) — do not average or bundle claims.

Claim Category	Example Assertion	Key Distinction to Apply
Stolen/rigged election	“The 2020 election was stolen” / “rigged”	Bare assertion is often opinion-adjacent; becomes a scorable factual claim when paired with a specific falsifiable mechanism (“machines switched votes,” “millions of illegal ballots”)
Voting-machine tampering	Dominion/ES&S machines altered vote counts	Vulnerability ≠ exploitation. Evidence a system could theoretically be manipulated is not evidence it was
Dead voters / illegal voters	“Thousands of dead people voted”	Requires a specific, checkable count from an authoritative source (state election officials, courts), not an extrapolated or unverified figure
Mail-in ballot fraud	Mass mail-in ballot fraud changed the outcome	Individual, prosecuted fraud cases (which exist at small scale in most elections) are not evidence of fraud at a scale that could change a statewide or national result
Foreign interference conflation	Documented foreign influence operations “prove” the outcome was altered	Influence effort ≠ impact. The declassified March 2021 Intelligence Community Assessment found foreign influence and interest, but no indication any foreign actor altered vote counts, registration, tabulation, or reporting
Certification illegitimacy	Officials “illegally” certified a “fraudulent” result	Analyst disagreement or process irregularity ≠ cover-up. Certification is a ministerial duty; refusal without evidence of fraud is itself a legal violation (see election-law-and-administration)
Agency politicization / cover-up	Intelligence or DOJ officials “covered up” the truth	Politicization inside an agency ≠ the election being rigged. Internal disagreement, leaks, or a documented politicized decision are separate claims from “votes were changed”
Verification/institution delegitimization	Fact-checkers, courts, or the press are “fake,” “corrupt,” or “in on it”	This is the Verification Delegitimization category from spokesperson-misinformation-assessment — attacks the correction mechanism itself rather than rebutting the substance; score as its own high-weight claim

---

Section 5: Factual Record Baseline (2020 Election)

Use this baseline as the evidentiary anchor for “what the evidence actually shows” rebuttals. Update it (Section 7 protocol) if analyzing claims about a different election cycle.

• The 2020 election was extensively audited, recounted, and litigated. Trump and allied campaigns filed 60+ post-election lawsuits alleging fraud; courts dismissed the overwhelming majority for lack of evidence, and several resulted in sanctions against filing attorneys (Giuliani, Eastman — see judicial-decision-analytics-specialist for docket-level detail).
• All 50 states certified their results, including states with Republican governors, secretaries of state, and legislatures.
• The intelligence community’s declassified March 2021 assessment found that foreign actors (Russia, Iran, and others) conducted influence operations, but found no indication that any foreign actor altered any technical aspect of voting — registration, ballot casting, tabulation, or reporting of results.
• State-level audits (including the Arizona “Cyber Ninjas” partisan audit) confirmed the certified outcome despite methodology criticized as unreliable.
• Multiple participants in the fake-elector scheme faced criminal charges in Michigan, Nevada, Georgia, Arizona, and Wisconsin, plus federal charges.

Cite election-law-and-administration for HAVA/NVRA/UOCAVA procedural framework and public-corruption-ombudsman §5 (“Election Denialism”) for the fuller evidentiary record and actor-level documentation.

---

Section 6: History of Election-Fraud Conspiracy Involvement (Speaker Research)

Before scoring a document, research the speaker’s documented history with election-fraud claims to establish pattern context (a factor in the EDS scoring below). Maintain a running summary; update via Tavily research (Section 8) rather than relying solely on static knowledge, since this record continues to develop.

Baseline Timeline (Trump)

Period	Documented Pattern
2016	Claimed the primary and general election were “rigged” before losing states were even final; won the Electoral College but continued to falsely claim millions of illegal votes cost him the popular vote; established the Presidential Advisory Commission on Election Integrity (disbanded 2018 without finding evidence of widespread fraud)
2020	“Stop the Steal” — sustained false claims the election was stolen via mail-in fraud, machine tampering, and dead/illegal voters; 60+ lawsuits filed and dismissed; pressure on state officials (e.g., the recorded call urging Georgia’s Secretary of State to “find” votes)
Jan 6, 2021	Fake-elector scheme across seven states; rally speech preceding the Capitol attack; later pardons/commutations for January 6 defendants
2021-2025	Continued repetition of the “stolen election” claim (a “Bottomless Pinocchio” pattern under spokesperson-misinformation-assessment‘s framework) across rallies, social media, and interviews, despite certification, audits, and dismissed litigation
2025-2026	Declassification-theater pattern (Section 3) — administration officials releasing selectively framed material described as new evidence of 2020 election fraud or foreign interference

Research protocol for updates: run Tavily searches such as "Trump" election fraud claims [year] fact check, "[official name]" election integrity declassification [year], and Trump 2020 election lawsuit dismissed sanctions to refresh this timeline before scoring a current document. Prioritize AP, Reuters, NPR, ProPublica, Brennan Center, and court records; treat single-outlet or partisan-source claims as unverified pending corroboration.

---

Section 7: The Gate — Protected Opinion vs. False Factual Claim

Run this gate before any scoring, for each individual claim tagged under Section 4. This reuses the gate-then-score architecture of spokesperson-misinformation-assessment and malice-evaluator.

The Boundary Test

“Is this a value judgment or characterization that a reasonable person could hold regardless of the facts — or does it assert a specific, checkable fact about how the election was conducted?”

Classification	Definition	Example
Protected Opinion	A value judgment, prediction, or characterization not reducible to a checkable factual assertion	“This was the most dishonest election in history” — contestable opinion, not a discrete fact
False Factual Claim	A specific, checkable assertion about election administration, contradicted by authoritative sources (courts, certified results, audits, the government’s own intelligence assessment)	“Voting machines were programmed to switch votes from Trump to Biden” — a specific, falsifiable claim rejected by every relevant audit and court proceeding
Borderline	Could be either depending on specificity and evidence — apply the five-factor test below	“The election had serious irregularities” — vague enough to be opinion unless paired with a specific debunked mechanism

The Five Distinguishing Factors

Any 3 of 5 tips a borderline claim to False Factual Claim:

1. Specificity — does the claim name a specific, checkable mechanism (a machine vendor, a vote count, a named county) rather than a vague impression?
2. Prior debunking — has this specific claim already been rejected by a court, an official audit, or an institutional fact-checker?
3. Repetition after correction — is the speaker repeating the claim after the debunking was public (the “Bottomless Pinocchio” pattern)?
4. Institutional-trust targeting — does the claim’s function target trust in courts, election administration, or the press itself, rather than argue a policy position?
5. Authoritative contradiction — does the claim directly contradict the government’s own findings (e.g., the IC’s March 2021 assessment) or the speaker’s own prior statements?

Advocacy Safe Harbor (NEVER scored as a false factual claim)

• Arguing for stricter election-integrity policy (voter ID, proof of citizenship, paper trails) on the merits
• Predicting that mail-in voting could be more vulnerable to fraud, as a policy argument, without asserting a specific event occurred
• Criticizing the speed, transparency, or communications of election administration
• Expressing distrust or disappointment in an outcome without asserting a specific false mechanism
• Reporting a genuine, documented vulnerability, isolated fraud case, or real (limited) foreign influence effort accurately and without conflating it with vote-count impact

Reader-Facing Transparency Output

Every scored claim must include:

**Protected Opinion vs. False Factual Claim Determination:** **False Factual Claim.**
[1-3 sentence explanation citing the specific distinguishing factors met, and what
authoritative source contradicts the claim.] The protected-opinion version of this
message would be: [1 sentence describing the value-judgment framing available to the
speaker without the false factual assertion].

---

Section 8: Election Doubt-Sowing Severity (EDS) Scoring

Every claim that passes the Gate as a False Factual Claim receives an Election Doubt-Sowing Severity (EDS) score, 1-5, structurally mirroring the MSS scale in spokesperson-misinformation-assessment.

Score	Label	Criteria	Evidence Required
1	Unsubstantiated Assertion	Factually false or unsupported claim, no evidence of prior knowledge of the correction	An authoritative source (court ruling, certified count, official audit) contradicting the claim; no evidence the speaker was previously corrected
2	Reckless Repetition	The correcting information (a court ruling, audit, certified result) was public before the statement was made	Above + documented public availability of the correction before this statement
3	Knowing Distortion	The claim contradicts the speaker’s own prior on-record acknowledgment, the government’s own intelligence/audit findings, or a final court ruling the speaker was a party to	Above + a specific contradicting document the speaker had direct access to (e.g., the March 2021 IC assessment, a ruling in a case the speaker’s campaign filed)
4	Directed Falsehood Campaign	The same debunked claim is repeated across 2+ venues (rally, social media, press conference, interview) after a formal correction was public	Above + dated evidence of repetition-after-correction across multiple venues
5	Systemic Doubt-Sowing Operation	A sustained, multi-claim, multi-channel pattern explicitly combined with institution-delegitimization (courts, press, election officials) over an extended period	Above + a documented multi-claim pattern with explicit institutional-trust targeting sustained over time (e.g., the 2020-2026 pattern in Section 6)

Evidence-Tier Gating

• Court ruling, certified result, or official audit exists and directly contradicts the claim: eligible for EDS 1-5
• Credibly reported (major outlet) but not yet adjudicated or audited: caps at EDS 2
• Only a general sense the claim “seems dubious,” no specific contradicting authority: caps at EDS 1 and cannot establish the “knowing” element required for EDS 3+
• No authoritative source contradicts the specific claim: not eligible — do not score; note as unresolved rather than assumed false

Document-Level Cumulative Designation

When a single document contains multiple EDS-scored claims, assign a document-level designation:

Designation	Threshold
Isolated Claim	1 claim scored, EDS 1-2
Doubt-Sowing Pattern	2+ claims scored at EDS 2+, OR 1 claim at EDS 4+
Sustained Doubt-Sowing Campaign	3+ claims scored at EDS 3+, OR any claim at EDS 5, OR a documented “repetition after correction” pattern within the same document

---

Section 9: Tavily Web-Research Protocol

Use Tavily Search (via TAVILY_API_KEY) for the three research tasks this skill requires. Never fabricate a transcript, quote, or fact-check citation — if a source cannot be located, say so explicitly and mark the claim unresolved.

Task 1: Locate the Transcript or Primary Document

Situation	Example Query
User references a speech by name/date but has not pasted the text	"[speaker]" speech transcript [date] [venue]
User references a declassification event	"[agency]" declassified [topic] [date] official statement
Confirming exact wording of a quote	"[exact phrase]" transcript site:rollcall.com OR site:c-span.org

Prefer official transcripts (White House, C-SPAN, Congress.gov, agency press releases) over paraphrased news summaries when quoting exact language.

Task 2: Verify Claims Against Institutional Fact-Checks

Situation	Example Query
A specific factual claim needs a fact-check citation	"[specific claim]" fact check PolitiFact OR FactCheck.org OR AP
Confirming a court ruling or audit result	"[case name or state]" election audit ruling result [year]
Checking the IC assessment or other official findings	intelligence community assessment 2020 election foreign interference declassified

Apply the same trusted-source hierarchy as public-corruption-ombudsman: accept official documents, AP/Reuters/NPR/major newspapers, PolitiFact/FactCheck.org/Snopes/WaPo Fact Checker, Brennan Center, court records; verify single-source or smaller-outlet claims against a second source; reject partisan blogs and unverified social media as standalone evidence.

Task 3: Update Speaker History

Per Section 6’s research protocol — run before scoring to confirm the pattern-context factor is current.

Provenance Tagging

Per .cursor/rules/patriot-investigative.mdc and the ITI inferred-data-transparency rule, tag every piece of AI-assisted output distinctly from sourced fact:

• (Human-verified) — a direct quote confirmed against an official transcript or primary document
• (AI-extracted) — a quote or claim pulled from a document by this skill’s own reading, not yet independently verified
• (AI-inferred) — a device tag, gate classification, or EDS score, all of which are this skill’s analytical judgment, not sourced fact
• (Estimated) — any approximate date, count, or figure not drawn from an authoritative source

---

Section 10: Defamation-Law Protection Framework

This skill reuses the legal-safety architecture of spokesperson-misinformation-assessment and malice-evaluator, since its subjects (a sitting or former president, cabinet officials, elected Republicans) are public officials or public figures under US defamation law.

Defense	How This Skill Implements It
Truth (Substantial Truth)	Every EDS-scored claim requires an authoritative contradicting source (court ruling, certified result, audit, official assessment) as its factual predicate. The predicate — “on [date], [court/audit/assessment] found X” — is literally true.
Fair Report Privilege	This skill analyzes public speeches, public documents, and public court/audit records — all privileged reporting.
Opinion Privilege	The device tags, the gate classification, and the EDS score are evaluative judgments about rhetorical technique and communicative function — protected opinion under Milkovich v. Lorain Journal (1990). The factual predicate (the contradicting authority) is cited separately from the evaluative conclusion.
Public Official / Public Figure Doctrine	Subjects must show actual malice (New York Times v. Sullivan, 1964) to prevail; the evidence-gated, source-cited methodology forecloses that showing.
Qualified Privilege	Civic education and accountability journalism serve a legitimate public interest.

Required Legal Disclaimer

Include this disclaimer in every article-ready output (Section 11, Format A):

### Legal Disclaimer

This analysis applies a published analytical framework — derived from the Institute for
Propaganda Analysis's seven propaganda devices — to documented public statements by a
public official or public figure. Each Election Doubt-Sowing Severity score rests on a
cited authoritative source (a court ruling, certified election result, official audit,
or government assessment) that contradicts the specific factual claim scored. The
severity score is an evaluative judgment about the claim's relationship to that record,
not a statement of criminal guilt or a finding that the speaker committed a crime.
*New York Times Co. v. Sullivan* and its progeny apply to this assessment of a public
official's or public figure's public statements.

---

Section 11: Output Format

Format A: Article-Ready Analysis

# Election Propaganda Analysis: [Document Title/Speaker/Date]

**Document:** [speech/press conference/post title, speaker, date, venue]
**Analyst:** Patriot University AI — Election Propaganda Ombudsman Skill
**Sources consulted:** [transcript source, fact-check sources]

## Executive Summary
[2-4 sentences: what the document claims, which devices dominate, overall doubt-sowing
pattern designation, and the topline "what the evidence shows" contrast.]

## Propaganda Device Analysis

### [Device Name]
**Quote:** "[exact text]"
**Watch for:** [tell that triggered this tag]
**Ask yourself:** [diagnostic question]
**Confidence:** [High/Medium/Low]

[Repeat for each device instance found]

## Declassification Theater Pattern
[Present only if applicable — describe how devices combine and note the speed-gap
mechanism.]

## Election Doubt-Sowing Claim Assessment

| # | Claim | Category (§4) | Gate | EDS | Contradicting Authority |
|---|-------|---------------|------|-----|--------------------------|
| 1 | [claim text] | [category] | [Protected Opinion / False Factual Claim] | [1-5 or N/A] | [source] |

### Claim 1: [short title]
**Protected Opinion vs. False Factual Claim Determination:** [per Section 7 template]
**EDS:** [score] — [label] (if scored)
**Fact-check / authoritative source:** [citation]

[Repeat per claim]

## What the Evidence Actually Shows
[Section 5 baseline facts relevant to the claims scored, stated plainly, with citations.]

## Document-Level Designation
**[Isolated Claim / Doubt-Sowing Pattern / Sustained Doubt-Sowing Campaign]** — [one
sentence justification citing the specific claims and scores].

## Pattern Context
[1-2 sentences connecting this document to the speaker's documented history from
Section 6, with citation to specific prior instances.]

### Legal Disclaimer
[Per Section 10]

## Sources
[Numbered citations for every quote, fact-check, court ruling, and audit referenced.]

Format B: Accountability Profile Feed

A compact, structured export block designed to be pasted into an accountability profile’s “Documented Actions” section (see accountability-profile-builder‘s template):

**Action:** Made false statements about election integrity in [venue] on [date]
**Date:** [date]
**Evidence:** [Election Propaganda Ombudsman analysis of (document title); EDS [score]
— [label]; contradicting authority: (citation)]
**Impact:** [effect on public trust / democratic process, stated plainly]
**Legal/Ethical Analysis:** [false factual claim contradicted by (authority); not
protected opinion because (distinguishing factors met)]

Include one such block per claim scored EDS 2+; omit claims that did not clear the gate.

---

Section 12: Cross-References

Skill	Relationship
patriot-speech-analyzer	General authoritarianism-marker detection across a broader taxonomy (executive overreach, judicial undermining, etc.); this skill is the election-integrity-specific specialization with a dedicated claim taxonomy and legal framework
spokesperson-misinformation-assessment	Provides the gate-then-score architecture and MSS scale this skill’s EDS scale mirrors; consumes this skill’s per-document findings when aggregating a spokesperson’s cumulative in-role record
malice-evaluator	Origin of the gate-then-score pattern (ideology vs. malice) and the defamation-law protection framework reused here
public-corruption-ombudsman	§5 “Election Denialism” is the primary actor-tracking home for documented election-fraud claims; this skill supplies the per-document technique analysis that ombudsman profiles cite as evidence
election-law-and-administration	Certification procedures, HAVA/NVRA/UOCAVA framework for evaluating “certification illegitimacy” claims
voter-suppression-law	Legal landscape for voter-ID and roll-purge policy claims, distinguishing legitimate advocacy from false factual assertion
voting-rights-act-expert	VRA case law context when a claim implicates discriminatory suppression, not just fraud narrative
election-threat-scoring	Consumes this skill’s document-level findings as evidence when scoring an actor’s election-threat composite score
judicial-decision-analytics-specialist	Docket-level detail on the 60+ dismissed 2020 lawsuits and any sanctions, cited as contradicting authority for EDS scoring
fact-checking	General claim-verification methodology; this skill applies a domain-specific evidence hierarchy and gate on top of it
source-credibility-scoring	Scores this skill’s own output’s sourcing rigor before publication — run after this skill, before patriot-sanity-check
citation-checker	Verifies citation accuracy in the article-ready output before publication
patriot-sanity-check	Runs last; validates that EDS scores and device tags are proportionate and evidence-based
patriot-editorial-framework	Governs how this skill’s findings are structured into a published Civic Accountability Format article
accountability-profile-builder	Consumes Format B output for the “Documented Actions” section of a profile

Board of Advisors

• pu-advisor-anne-applebaum — historian of authoritarianism and propaganda; consult for framing the broader pattern of information-control tactics
• pu-advisor-timothy-snyder — historian of 20th-century political atrocity and disinformation; consult for comparative-authoritarianism framing
• pu-advisor-liz-cheney — former Republican congresswoman and January 6 investigator; consult for the Republican-establishment and election-subversion angle

---

Section 13: Guardrails and Anti-Patterns

What This Skill Is NOT

• NOT a partisan opinion scorer — it never scores protected political opinion, only specific, checkable false factual claims
• NOT a criminal accusation tool — an EDS score is an evaluative judgment about a claim’s relationship to the documented record, not a finding of criminal intent
• NOT applicable to policy advocacy — arguing for election-integrity legislation on the merits is always Protected Opinion / Advocacy Safe Harbor
• NOT a substitute for accountability-profile-verification — profile-level evidence sourcing still runs its own verification pass
• NOT a one-size-fits-all speech analyzer — its claim taxonomy (§4) is calibrated to the “stolen/rigged election” narrative pattern; for unrelated rhetoric, use patriot-speech-analyzer

Hard Guardrails

1. The Gate runs before any scoring — no EDS score without a Protected-Opinion-vs-False-Factual-Claim classification first.
2. Authoritative contradiction required for EDS 1+ — no court ruling, certified result, audit, or official assessment contradicting the specific claim means no score; note as unresolved.
3. Knowledge must be documented for EDS 3+ — a general “this seems false” judgment caps at EDS 1.
4. Advocacy Safe Harbor is absolute — genuine policy argument, even one-sided, is never scored as a false factual claim.
5. Never fabricate a quote, transcript, or fact-check citation — if Tavily research cannot locate a primary source, say so explicitly rather than approximating.
6. Device tags are analytical judgments — always disclosed as (AI-inferred), never presented as an established fact independent of the underlying quote.
7. Sanity-check composition — patriot-sanity-check runs last and may flag disproportionate scores or unsupported device tags.

Common Mistakes to Avoid

Mistake	Correction
Scoring “the election was unfair” as a false factual claim	It is a value judgment absent a specific mechanism — Protected Opinion
Treating a real, documented vulnerability as proof of exploitation	Vulnerability ≠ exploitation — cap at EDS 1 unless exploitation is separately documented
Treating documented foreign influence efforts as proof of vote-count impact	Influence effort ≠ impact — the March 2021 IC assessment found no vote-tampering; cite it
Inferring “knowing” falsity from a single false statement	A false claim alone caps at EDS 1-2; EDS 3+ needs documented prior knowledge
Scoring a document with no device tags found	Absence of the seven devices is a valid finding — report “no significant propaganda techniques detected” rather than forcing a tag
Publishing a raw EDS score without the reader-facing determination text	Every scored claim requires the transparency block (Section 7)

---

Section 14: Worked Example — Declassification Theater

Input: A presidential address announcing “really big” newly declassified intelligence about 2020 election voting-machine vulnerabilities, foreign interference, and voter rolls, delivered with agency officials present.

Device analysis:

• Transfer (High confidence): the podium, agency officials, and “declassified” framing borrow institutional authority for claims not yet independently evaluated
• Testimonial (High): officials “vouching” for the conclusion without showing underlying evidence
• Card-Stacking (High): selective release framed as proof of a sweeping conclusion
• Glittering Generalities (Medium): “election integrity” and “transparency” invoked without a specific, checkable claim in that instance
• Plain Folks (Medium): framing the release as being “for the American people,” obscuring that the speaker controls the releasing agency
• Bandwagon (Medium, requires post-speech monitoring): rapid amplification across allied media within minutes, before independent review is possible

Declassification Theater pattern: flagged — the speed gap between the announcement and any independent assessment of the underlying documents is the central mechanism.

Claim gate example: “This proves the 2020 election was rigged by foreign actors” → Boundary test: is this a checkable mechanism claim? Yes — asserts foreign actors changed the outcome. Factors met: authoritative contradiction (the March 2021 IC assessment found no vote-tampering — factor 5), repetition after correction if this claim has been made before (factor 3, pending Section 6 research). Classification: False Factual Claim.

EDS scoring: if the speaker has made materially the same claim before, after the IC assessment was public → EDS 3 (Knowing Distortion) at minimum, per Section 8’s evidence table; escalate to EDS 4 if repeated across 2+ venues after this address.

Document-level designation: if 2+ such claims clear the gate at EDS 2+, designate Doubt-Sowing Pattern; if the address combines this with explicit attacks on courts, auditors, or fact-checkers as part of the same speech, escalate toward Sustained Doubt-Sowing Campaign pending the multi-claim threshold in Section 8.