Investigation Lifecycle

Argus follows a staged lifecycle from intake through completed output.

01

Intake

Submitted documents, images, URLs, names, and case notes are stored and queued for processing.

02

Extraction

OCR, text extraction, and claim-oriented parsing turn raw files into extracted text, entities, and candidate claims.

03

Planning

The system identifies target entities, likely verification lanes, registry routes, and bounded research tasks.

04

Research

Argus executes screening, website/domain checks, registry routing, Stage 3A research, and enhanced web tasks where applicable.

05

Evidence Normalization

Findings are converted into structured evidence with source tiering, treatment limits, and contradiction posture.

06

Projection

The report projection plan maps evidence into sections such as Claim Verification, Location Intelligence, Risk Register, and Source References.

07

Report Rendering

Argus builds the enhanced report view model, memo view, and completed-case artifacts from the deterministic evidence set.

08

QA and Export

Quality checks validate label hygiene, source references, omission handling, answer coverage, and output safety before export or review.

Runtime truth

The orchestration entry point is server/investigationRunner.ts, while the completed enhanced report is rebuilt on read through the analysis/report path rather than stored as the single persisted truth artifact.