---
module_id: KB-00
version: 0.4.2
status: baseline_draft
edition: public
language: en
role: navigation
updated_on: 2026-09-06
requires: []
---

# Editorial AI Standard Knowledge Base: Index

**Short name:** EAS-KB  
**Version:** 0.4.2  
**Scope:** Platform-independent baseline for deriving editorial AI prompts, knowledge packs, application policies, output contracts, runtime controls and evaluation tests.

EAS-KB is a starting framework, not a universal or complete newsroom standard. It should be adapted to the relevant newsroom, language, jurisdiction, audience, workflow and risk profile. The baseline does not itself authorize publication or external action.

## Current canonical modules

| ID | File | Canonical responsibility |
|---|---|---|
| KB-00 | `00-KB-index.md` | Navigation, scope and module map |
| KB-01 | `01-KB-principles-and-rule-hierarchy.md` | Principles and rule hierarchy |
| KB-02 | `02-KB-editorial-judgment-and-input-analysis.md` | Editorial judgment and input analysis |
| KB-03 | `03-KB-sources-and-research.md` | Sources, research and evidence suitability |
| KB-04 | `04-KB-verification.md` | Claim verification, work states, verdicts and confidence |
| KB-05 | `05-KB-fairness-independence-and-public-interest.md` | Fairness, right of reply, independence and public interest |
| KB-06 | `06-KB-privacy-harm-and-sensitive-content.md` | Privacy, identification, harm and sensitive content |
| KB-07 | `07-KB-corrections-and-accountability.md` | Corrections, updates, complaints and accountability |
| KB-08 | `08-KB-ai-editorial-governance.md` | Editorial AI governance, confidentiality, synthetic media and authority |
| KB-13 | `13-KB-workflows-output-control-and-escalation.md` | Task modes, workflow, output control and escalation |
| KB-90 | `90-KB-application-derivation.md` | Deriving scoped applications from the baseline |
| KB-91 | `91-KB-sources-and-decisions.md` | External references and design basis |
| KB-92 | `92-KB-tests-and-release-policy.md` | Testing, versioning and release gates |

## Planned or incomplete specialist modules

| ID | Planned file | Scope |
|---|---|---|
| KB-09 | `09-KB-writing-and-editing.md` | Writing and editing methodology |
| KB-10 | `10-KB-enrichment-and-explanation.md` | Evidence-based enrichment and explanation |
| KB-11 | `11-KB-formats-and-output-patterns.md` | Output structures and format-specific patterns |
| KB-12 | `12-KB-language-russian.md` | Russian language profile |
| KB-14 | `14-KB-quality-assurance.md` | Quality assurance and optional scoring |

A planned or incomplete module has no operative authority until its content is included in a release and selected by the derived application.

## Suggested reading paths

### General editorial application
Use KB-01, KB-13 and KB-02. Add KB-03 and KB-04 when sources or factual verification matter. Add KB-05 through KB-08 when publication responsibility, privacy, corrections or AI governance are relevant.

### Fact-checking application
Use KB-01, KB-13, KB-03 and KB-04. Add KB-02 for editorial significance and input analysis. Include KB-05, KB-06 and KB-07 when the output may be published or may affect identifiable people.

### Source-bound editing or translation
Use KB-01, KB-13 and KB-02. Add KB-05 and KB-06 when the transformation can alter fairness, identification, privacy or sensitive meaning. External research should only be activated when the task or application profile requires it.

### Research or monitoring
Use KB-01, KB-13, KB-02, KB-03 and KB-04 as applicable. Add publication-responsibility modules before consequential public output or disclosure.

### Application construction
Use KB-90 together with the relevant capability modules. Record the EAS-KB version, selected rules, local profiles, dependencies, output contracts, tools, permissions and tests.

## Canonical ownership rule

Each substantive norm should have one canonical owner. Other modules may activate or reference that norm, but should not redefine it independently. A derived application should include enough dependency context for selected rules to remain operational.

## Machine-readable artifacts

The public edition includes:

- `../governance/rule-registry.json` for rule IDs, ownership, levels and test references;
- `../governance/tests.json` for the current test inventory and execution state;
- `../governance/source-registry.json` for external reference sources;
- `../examples/evidence-record-contract.json` for the verification record contract;
- `../schemas/` for example application, newsroom and derivation schemas;
- `../tools/validate_public_package.py` for deterministic package checks.

These artifacts support reproducibility. They do not replace editorial judgment or behavioral testing.

## Status boundary

`baseline_draft` means that the material is intended as a reusable starting point. It does not mean that the framework is complete, independently certified, legally sufficient for every jurisdiction, approved by a specific newsroom, or proven safe for a particular AI application.