Argus
Internal intelligence system documentation
A branded internal knowledge base for analysts, engineers, and future collaborators working on investigation pipelines, Mission Control, Stage 3A research, deterministic report generation, MiniMax preview policy, and live QA operations.
Argus uses structured extraction, bounded research, evidence normalization, projection planning, quality checks, and controlled preview-only experiments to produce sourced due-diligence outputs.
Docs map
Start Here
Orient new analysts and engineers around what Argus is, what it does, and how the platform is meant to be used.
Investigation Lifecycle
Follow the path from intake and extraction through evidence, projection, report generation, and QA.
Enhanced DD Reports
Understand the deterministic report model, section semantics, source references, and verification requirements.
Stage 3A Research
Review bounded web research planning, triage, fetch, extraction, and mapping back to findings.
Mission Control
Understand the live investigation surface, status lanes, location intelligence, and event stream.
Runbooks
Use smoke tests, debug traces, fixture workflows, and release checklists to operate the system safely.
QA Matrix
Study the standing regression fixtures that protect projection quality, location handling, and case-specific edge conditions.
Policies
Review the rules around evidence use, adverse promotion, MiniMax previewing, imagery handling, and redaction.
Operational shortcuts
pnpm smoke:enhanced-dd-v2 -- --case=bluetriangle --require-vision --require-web --require-pdf
pnpm test:location-previews
pnpm smoke:minimax-preview