Scoring & Inclusion Systems — Complete Inventory
All scoring rubrics, inclusion gates, severity classifications, and quality-assurance checks applied to Patriot University documents, profiles, and content — spanning Skills, Agents, and Cursor Rules.
Skills
Accountability Profile Pipeline
Runs in strict composition order. Each step must complete before the next begins.
| Step | Skill | What It Does |
|---|---|---|
| 0 | accountability-profile-builder |
End-to-end authoring workflow. Defines the P0/P1/P2/Media priority criteria, the Documented/Credibly Reported/Alleged/Excluded evidence tiers, and the prohibited-language substitution table that all downstream gates rely on. |
| 1 | patriot-private-citizen-inclusion-gate |
Multi-factor eligibility test before any profile is created. Applies the Subject Classification framework (Public Official / Voluntary Public Figure / Limited-Purpose Public Figure / Private Citizen), the five Non-Speech Anchor criteria (A–E), the Speech Taint Strip Test, and requires a “Basis for Inclusion” disclosure block. Prevents inclusion solely on protected speech, political affiliation, or rally attendance. |
| 2 | accountability-profile-verification |
Evidentiary standards after eligibility is confirmed. Covers primary-source research, citation requirements, court record formats, removal of disallowed taxonomy sources, and legal-exposure risk assessment. Also enforces the J1–J5 Federal Judge Inclusion Gate for Article III judges. |
| 3 | malice-evaluator |
The Democratic Malice Assessment (DMA) — structured ideology-vs-malice gate (3-of-5 distinguishing factors) with a 1–5 DMS severity score and cumulative actor-level designation. Required for all trump-administration profiles and any P0/P1 profile. |
| 4 | patriot-sanity-check |
Meta-auditor that runs last. Runs a Level 1–5 audit (factual claims, proportionality, legal claims, systemic consistency, hostile-scrutiny stress test) and issues an overall verdict (PASS / PASS WITH CAVEATS / NEEDS REVISION / FAIL). |
Election & Democratic Health Scoring
| Skill | What It Does |
|---|---|
election-threat-scoring |
Multi-dimensional threat matrix scoring actors, state vulnerabilities, and federal threats on Likelihood × Impact × Urgency (each 1–5). Maps to NIST-style Prevent/Prepare/Mitigate/Respond/Recover actions. |
election-threat-scenario-planner |
Peter Schwartz 8-step scenario planning applied to election threats — produces prioritized scenario narratives with leading indicators, not a numeric score. |
democratic-health-monitoring |
V-Dem, Freedom House, and Protect Democracy indicator sets with severity scores and escalation velocity analysis. |
patriot-speech-analyzer |
Authoritarianism marker detection in political rhetoric, with per-marker severity assessment and confidence calibration. |
Content & Source Quality Scoring
| Skill | What It Does |
|---|---|
news-credibility-scoring |
Multi-dimensional credibility scoring for news sources — ownership analysis, editorial standards, correction history, domain authority. Produces tier classification. |
claims-integrity-audit |
Audits ITI marketing/portfolio documents for false, hyperbolic, or unverifiable claims. Classifies each claim as Puffery / Objective-verifiable / Comparative / Capability with a verdict and remediation. |
fact-checking |
Professional fact-checking methodology with a four-tier source credibility hierarchy (Tier 1 = .gov/AP/Reuters → Tier 4 = blogs/social). |
citation-checker |
Legal-citation verification. Reads citator/Shepard’s signals (Red flag / Yellow flag / Green signal) and classifies output to one of three Quality Tiers — Tier 1 Court-ready, Tier 2 Research-grade, Tier 3 Educational — depending on intended use. |
patriot-source-scanner |
RSS/Tavily content acquisition with source-category priority (High = wire services / official statements; Medium = major outlets / press briefings; Low = analysis / opinion). Priority drives scan cadence, not content credibility. |
Investigative Context (Scoring-Adjacent)
| Skill | What It Does |
|---|---|
trump-corruption-accountability-tracker |
Tracks corruption patterns across self-dealing, crypto, foreign influence, retribution. Evidence-based narrative — feeds profile updates but is not a scored rubric. |
public-corruption-ombudsman |
Comprehensive accountability registry manager — surfaces actors, documents patterns, prepares truth-and-reconciliation evidence from 293+ profiles. |
Legislative & Policy Impact Scoring
Distinct from the profile pipeline above: this family scores a piece of legislation or executive action, not a person.
| Skill | What It Does |
|---|---|
legislative-impact-scoring |
The Legislative Impact Assessment (LIA) — scores a single federal bill, executive order, agency rule, or enforcement change on two parallel tracks that are never combined into one composite number (the non-collapse rule / knob 22). A Q-track (fiscal / distributional / sectoral) aggregates-and-attributes others’ estimates (CBO/JCT/TPC/Tax Foundation/PWBM/Yale) with all 22 contestable “knobs” disclosed; an R-track (rights-exposure / burden-chilling / institutional severity) runs the doctrinal-legal workflow by composing the existing legal-expert and democratic-health-monitoring skills and reports doctrinal exposure + litigation status, never an unconstitutionality verdict. Shared back-half: provenance gate, implementation, ex-post/realized (defaults “too-early”), and a bipartisan orientation-paired divergence layer. Governed by docs/legislative-impact-methodology.md; the 22 knobs + source reference class live in docs/legislative-impact-knob-baseline.md. PolicyEngine original estimation is a labeled Phase 2 deferral; the live publish path (scorecard rendering, KB taxonomy home) is deferred per docs/legislative-impact-checkpoint.md. |
Agents
| Agent | What It Does |
|---|---|
News Agent (patriot-university-news-agent) |
Orchestrates source scanning → content analysis → threat score aggregation → profile updates on a scheduled basis. |
Patriot News Orchestrator (patriot-news-orchestrator) |
Coordinates the Scan → Analyze → Score → Update pipeline. |
| Malice Evaluator Agent | Runs DMA scoring on profiles as a dedicated agent. |
| Claims Ombudsman Agent | Audits documents for false/misleading claims; gates publication. |
| Claims Evidence Curator Agent | Maintains evidence registries and runs staleness checks on claims. |
Rules (.cursor/rules/*.mdc)
| Rule File | Trigger Scope | What It Enforces |
|---|---|---|
patriot-malice-evaluator.mdc |
knowledgebase/accountability/**/*.md |
Automatically triggers the malice-evaluator skill. Defines the 4-step composition order (inclusion gate → verification → DMA → sanity check). Enforces that all trump-administration profiles have either a DMA designation or an explicit “No DMA” transparency block. |
patriot-accountability-profile-standards.mdc |
knowledgebase/accountability/**/*.md |
Evidence Tier system (§1), Criminal Activity Claims language standards (§2), Citation format (§3), P0/P1/P2 priority criteria (§4), Speech Is Not a Crime standing principle (§5), Private Citizen Inclusion Gate with 5 Non-Speech Anchors (§5A), Basis for Inclusion disclosure block requirement (§5B), Federal Judge Inclusion Gate J1–J5 (§5C), Factcheck/correction mechanism (§6), Education Only profile extra-scrutiny rule (§7), Legal Exposure Guardrails (§8), Verification Workflow (§9), Prohibited Language list (§10). |
Score & Gate Type Reference
| Type | System | Scale / Output |
|---|---|---|
| Subject classification | patriot-private-citizen-inclusion-gate |
Public Official / Voluntary Public Figure / Limited-Purpose Public Figure / Private Citizen |
| Inclusion eligibility (private citizen) | patriot-private-citizen-inclusion-gate |
Pass / Fail against 5 Non-Speech Anchors (A–E) + Speech Taint Strip Test |
| Inclusion eligibility (federal judge) | patriot-accountability-profile-standards.mdc §5C |
Pass / Fail against the J1–J5 Federal Judge Inclusion Gate |
| Evidence tier | patriot-accountability-profile-standards.mdc §1 |
Documented → Credibly Reported → Alleged → Excluded |
| Profile priority / severity | patriot-accountability-profile-standards.mdc §4 + accountability-profile-builder |
P0 / P1 / P2 / Media |
| Democratic Malice Score (DMS) | malice-evaluator skill |
1–5 per action, with cumulative actor-level designation |
| Ideology vs. Malice classification | malice-evaluator skill |
Ideology / Malice (3-of-5 distinguishing factors gate) |
| Sanity-check audit verdict | patriot-sanity-check skill |
PASS / PASS WITH CAVEATS / NEEDS REVISION / FAIL (across 5 audit levels) |
| Election composite threat score | election-threat-scoring skill |
(Likelihood × Impact × Urgency) / 25 normalized to 1–5, mapped to Critical / High / Elevated / Guarded / Low |
| State vulnerability tier | election-threat-scoring skill |
Tier 1 Critical (4.0+) → Tier 5 Low (<1.0) |
| Authoritarianism marker severity | patriot-speech-analyzer skill |
Per-marker severity + confidence level |
| Democratic health severity | democratic-health-monitoring skill |
5-level severity scale (Level 1 Healthy → Level 5 Authoritarian); Freedom House 0–100; V-Dem sub-indices; 90-day velocity index |
| Legislative Impact Assessment (LIA) | legislative-impact-scoring skill |
Two-track, never combined: Q-track (attributed fiscal/distributional/sectoral estimates + 22 disclosed knobs) and R-track (rights-exposure direction + doctrinal exposure settled/contested/likely-unconstitutional + burden-chilling + democratic-health severity). Scores legislation/executive action, not a person. |
| Source credibility tier (news) | news-credibility-scoring skill |
Tier 1–4 + composite score |
| Source credibility tier (fact-check) | fact-checking skill |
Tier 1 (.gov/AP/Reuters) → Tier 4 (blogs/social) |
| Citation quality tier | citation-checker skill |
Tier 1 Court-ready / Tier 2 Research-grade / Tier 3 Educational |
| Source scan priority | patriot-source-scanner skill |
High / Medium / Low (drives scan cadence) |
| Claims integrity verdict | claims-integrity-audit skill |
Per-claim verdict (Accurate / Exaggerated / Unsupported / False) |
P0 / P1 / P2 Severity Definitions
Defined in patriot-accountability-profile-standards.mdc. Priority must be anchored to documented evidence, not network affiliation or political position alone.
- P0 — Documented direct participation in constitutional violations, seditious conspiracy, or fraud in office. Requires a court record or congressional finding. DMA scoring mandatory.
- P1 — Documented pattern of conduct that materially advanced democratic erosion (e.g., signed false certification, directed vote suppression, convicted of related crime). DMA scoring mandatory.
- P2 — Network-level amplifiers, political operatives, and enablers with documented but not individually criminal roles. DMA scoring recommended if ideology-malice boundary is in question.
DMS (Democratic Malice Score) Scale
Defined in the malice-evaluator skill. Scores ideology-vs-malice on a 1–5 rubric:
| DMS | Designation | Meaning |
|---|---|---|
| 1 | Ideological | Actions consistent with sincerely held political ideology; no evidence of intent to subvert democratic mechanisms |
| 2 | Opportunistic | Actions that exploited democratic weaknesses for personal or political gain without specific intent to subvert |
| 3 | Reckless | Actions that a reasonable actor would recognize as damaging to democratic mechanisms, regardless of stated intent |
| 4 | Purposeful | Actions taken with apparent intent to undermine specific democratic mechanisms |
| 5 | Systemic | Documented, sustained campaign to subvert multiple democratic mechanisms with evidence of specific intent |
Ideology vs. Malice Gate (3-of-5 Distinguishing Factors)
Runs FIRST in the malice-evaluator skill, before any DMS score is assigned. Each action is classified as Ideology (legitimate policy, no score) or Malice (scoring eligible). In the borderline zone, any three of five distinguishing factors tip the classification to Malice:
- Process subversion — Did the actor bypass normal process, oversight, or transparency requirements?
- Targeted asymmetry — Did the action disproportionately burden a specific protected group or political opponent?
- Expert rejection — Did the actor ignore expert, nonpartisan, or career-staff warnings?
- Pattern context — Has the actor or institution previously been enjoined, reversed, or sanctioned for similar conduct?
- Stated intent — Does the documented record (statements, communications, memos) show explicit intent to undermine democratic mechanisms?
Actions passed through legitimate legislative, executive, or judicial process are always classified as Ideology regardless of policy harm.
DMA Cumulative Actor-Level Designation
When an actor accumulates multiple qualifying actions (DMS 1+), they earn a cumulative designation. Defined in the malice-evaluator skill, Section 5:
| Designation | Threshold | Meaning |
|---|---|---|
| No DMA designation | Fewer than 2 qualifying actions, or all scores at DMS 1 | Actions may harm democracy but do not demonstrate pattern or clear intent |
| Pattern of Reckless Conduct | 2+ actions at DMS 3 (or any single DMS 4) | Pattern of conduct a reasonable actor would recognize as damaging |
| Pattern of Purposeful Subversion | 2+ actions at DMS 4 (or any single DMS 5) | Pattern showing apparent intent to undermine democratic mechanisms |
| Documented Systemic Campaign | 2+ actions at DMS 5 | Sustained, intentional campaign across multiple mechanisms |
DMS 3+ requires Tier 1 (Documented) evidence. Alleged-tier claims are not eligible for scoring.
Subject Classification Framework (Inclusion Gate)
Defined in the patriot-private-citizen-inclusion-gate skill, Section 1. Determines which inclusion threshold applies before any private-citizen profile is drafted:
| Classification | Criteria | Inclusion Threshold |
|---|---|---|
| Public Official | Sitting or former elected official; Senate-confirmed appointee | Standard accountability standard — documented official conduct |
| Voluntary Public Figure | Voluntarily assumed a prominent role (organized events, gave media interviews, made public claims of leadership) | Intermediate threshold; conduct-based anchor required but lower bar than private citizen |
| Limited-Purpose Public Figure | Voluntarily thrust into a specific public controversy by their own public actions | Intermediate threshold; must pass Non-Speech Anchor Test |
| Private Citizen | Has not voluntarily assumed a public role; known primarily through private associations, speech, or attendance | Highest threshold; must pass Non-Speech Anchor Test with Tier 1/2 evidence; priority capped at P1 absent a criminal conviction |
Non-Speech Anchor Criteria (A–E)
A private citizen or limited-purpose public figure may be included only if they independently satisfy at least ONE of the following five anchors, documented with Tier 1 or Tier 2 evidence. Defined in patriot-accountability-profile-standards.mdc §5A and the patriot-private-citizen-inclusion-gate skill:
| Anchor | Qualifying Conduct | Evidence Required |
|---|---|---|
| A | Criminal conviction, indictment, or formal charge related to the subject matter | Court record |
| B | Documented founder/officer/director/chapter-leader/operational-coordinator role in an organization with documented harmful conduct | Organizational filings, official records |
| C | Documented significant, targeted financial support that materially enabled specific harmful conduct | FEC filings or court record |
| D | Voluntarily and publicly assumed a leadership/coordination role via press conferences, media interviews, or public organizing (not merely attending events or posting online) | Tier 2 journalism specifically documenting the role |
| E | Used a documented official non-elected capacity (election official, law enforcement, government contractor) to advance the harmful conduct | Official records, court findings, congressional testimony, or Tier 2 journalism |
Never sufficient on their own: political positions, political affiliation, social media posts, rally attendance, organizational membership without documented leadership, or being listed in any external taxonomy (One6Project, SPLC, ADL, etc.).
Speech Taint Strip Test
A diagnostic applied during the inclusion gate: remove all speech and social-media evidence from the case. If only journalism, commentary, or expressive activity remains, the test fails and the subject cannot be included. If non-speech conduct (documented financial relationship, organizational role established by official records, direct coordination documented in court records) remains, the anchor may be met.
Federal Judge Inclusion Gate (J1–J5)
Defined in patriot-accountability-profile-standards.mdc §5C. Article III federal judges (district, circuit, Supreme Court) may only receive a personal accountability profile if the evidence record satisfies at least one of:
| Gate | Qualifying Conduct |
|---|---|
| J1 | Documented ex parte communications with a party or that party’s political operation outside the courtroom |
| J2 | Undisclosed financial interests or benefits connected to a party before the court — judicial financial disclosure violation |
| J3 | Personal corruption: gifts, travel, or benefits from parties or interests with matters before the court (ProPublica/Justice Thomas reporting standard) |
| J4 | Formal finding of judicial misconduct by the relevant Circuit Judicial Council under the Judicial Conduct and Disability Act (28 U.S.C. §§ 351–364) |
| J5 | Impeachment proceedings initiated by the House of Representatives |
Never sufficient on their own: erroneous rulings (even unanimously reversed), rulings that benefit the appointing president, being a particular president’s appointee, patterns of case management absent a Judicial Council finding, mere judicial complaints, or minority/unpopular legal theories.
When a judge’s rulings have significant structural impact but the personal-misconduct threshold is not met, the appropriate format is a Judicial Impact Analysis document (category: legal) — see aileen-cannon-judicial-impact.md as the precedent and deliberation record.
Evidence Tier System
Defined in patriot-accountability-profile-standards.mdc §1. Every factual claim in any accountability profile must be labeled with — and trace to — one of these four tiers:
| Tier | Definition | Labeling Rule |
|---|---|---|
| Documented | Primary sources: court records, indictments, verdicts, official congressional records, FEC filings, financial disclosures, signed executive orders | No label needed — it IS the claim |
| Credibly Reported | Multi-outlet journalism with named sources, published by recognized news organizations (AP, Reuters, NYT, WaPo, WSJ, NPR, ProPublica, etc.) | No label if multi-outlet corroborated |
| Alleged | Single-source claim, anonymous sourcing, or not yet independently confirmed | Prefix: “Allegedly,” or “According to [single source],” |
| Excluded | Unsourced assertions, AI inference, think-tank taxonomies without primary backing, anonymous blog posts, opposition-research databases without independent verification | Must NOT appear in profiles |
One6Project is an excluded source. If it was the sole basis for a person’s inclusion, the profile must either be substantiated with independent Tier 1/2 sources or be removed.
P0 / P1 / P2 / Media Priority — Cross-Reference
The accountability-profile-builder skill adds a fourth classification used to scope coverage of media personalities. The base rule is defined in patriot-accountability-profile-standards.mdc §4.
| Priority | Criteria | Notes |
|---|---|---|
| P0 | Documented direct participation in constitutional violations, seditious conspiracy, or fraud in office. Court record or congressional finding required. | DMA scoring mandatory |
| P1 | Documented pattern of conduct that materially advanced democratic erosion. | DMA scoring mandatory |
| P2 | Network-level amplifiers, political operatives, and enablers with documented but not individually criminal roles. | DMA scoring recommended in borderline cases |
| Media | Fox News hosts and other media personalities. Not P0/P1 unless specific documented non-journalistic conduct exists (direct coordination with coup plotters, financial conflicts, false statements under oath). | Opinion journalism, even deliberately misleading opinion journalism, is generally protected speech |
Private citizens are capped at P1 absent a criminal conviction; P0 is reserved for sitting or former officials with documented constitutional violations.
Sanity Check 5-Level Audit
Defined in patriot-sanity-check. Run last in the accountability profile pipeline. Each level is a discrete pass; the overall audit verdict aggregates the level results.
| Level | Audit Focus |
|---|---|
| Level 1 | Factual Claims Verification — Does each claim trace to a source meeting Tier 1–2 credibility? |
| Level 2 | Characterization Proportionality — Do severity/priority ratings match the documented evidence (not overclaimed, not underclaimed)? |
| Level 3 | Legal and Constitutional Claims — Are legal characterizations technically accurate? |
| Level 4 | Systemic Consistency — Are P0/P1/P2 classifications applied consistently across comparable subjects? |
| Level 5 | Hostile-Scrutiny Stress Test — Would the profile survive adversarial review under defamation challenge or political attack? |
Overall verdict: PASS / PASS WITH CAVEATS / NEEDS REVISION / FAIL. A FAIL blocks publication.
Citation Quality Tiers
Defined in the citation-checker skill. Output tier depends on intended use:
| Tier | Standard | Applies To |
|---|---|---|
| Tier 1 — Court-ready | Full Bluebook compliance; citator-verified; pinpoint-checked; quotation-verified | Appellate briefs, motions, court filings |
| Tier 2 — Research-grade | Existence verified; current authority confirmed; holdings accurately stated | Research memoranda, knowledge base skills |
| Tier 3 — Educational | Existence verified; holdings accurately stated; simplified citation format permitted | Know-your-rights materials, community education |
Citator/Shepard’s signals applied during verification: Red flag / Red stop sign (overruled or superseded — STOP); Orange Q / Yellow flag / Yellow triangle (negative or questioning treatment — investigate); Green signal (no negative treatment).
Election Threat Score — Composite Formula and Tiers
Defined in election-threat-scoring. Three independent 1–5 dimensions feed a normalized composite.
Composite Threat Score = (Likelihood × Impact × Urgency) / 25, normalized to a 1–5 scale.
| Composite Score | Threat Level | Action Priority |
|---|---|---|
| 4.0+ | Critical | Immediate, dedicated resourcing |
| 3.0 – 3.9 | High | Active mitigation |
| 2.0 – 2.9 | Elevated | Active preparation |
| 1.0 – 1.9 | Guarded | Monitoring |
| < 1.0 | Low | Baseline awareness |
State Vulnerability Tiers (May 2026 baseline)
| Tier | Score Range | Example States |
|---|---|---|
| Tier 1 — Critical | 4.0+ | GA, AZ, TX, FL |
| Tier 2 — High | 3.0 – 3.9 | OH, NC, WI, PA, NV, IN, MT |
| Tier 3 — Elevated | 2.0 – 2.9 | MI, NH, SC, LA, AL, AR, KS, TN, KY |
| Tier 4 — Guarded | 1.0 – 1.9 | VA, CO, NM, MN, ME |
| Tier 5 — Low | < 1.0 | CA, WA, OR, NY, IL, MA, CT, VT, HI |
Three scoring levels run independently: Level 1 Actor, Level 2 State Vulnerability, Level 3 Federal. Each produces its own composite that feeds the Prevent / Prepare / Mitigate / Respond / Recover action framework.
Democratic Health Severity (5-Level)
Defined in democratic-health-monitoring. Each indicator across the electoral, civic-liberties, accountability, and civic-life domains is rated on a 5-level severity scale.
| Level | Label | Description |
|---|---|---|
| 1 | Healthy | Indicator operating within established democratic norms |
| 2 | Cautionary | Localized or rhetorical strain; no systemic erosion |
| 3 | Eroding | Measurable degradation; pattern emerging across cases |
| 4 | Critical | Systemic degradation; institutional checks failing |
| 5 | Authoritarian | Institution captured or repurposed against democratic function |
Scoring rules: Score on actions and outcomes, not rhetoric alone (rhetoric is capped at Level 2 unless paired with action). Any score above 2 requires at least 3 corroborating sources. Distinguish isolated incidents from systemic patterns. Score changes require explicit evidence-cited justification.
Velocity index: Rate of change in severity scores over a rolling 90-day window; sustained acceleration is itself a leading indicator.
Education Only Profile — Extra Scrutiny Rule
Defined in patriot-accountability-profile-standards.mdc §7. Profiles containing the investigative-trail pointer block (the phrase Education only — verify independently. Absence of hits is not proof.) are flagged as requiring independent verification of every negative claim.
For these profiles:
- All negative claims must be sourced to Tier 1 or Tier 2 evidence.
- If One6Project (or a similar excluded source) was the only source, the negative claim must be removed or reclassified as “alleged pending independent verification.”
- The investigative trail block itself is appropriate and should be retained — it documents research pathways, not findings.