Intelligence-Research-OS

wiki/overview.md

Intelligence Research OS

Status: Local-first workbench foundation implemented; live UI-to-engine execution pending
Core engine: ResearchModule

Definition

Intelligence Research OS is the product layer built on top of ResearchModule.

ResearchModule is the file-gated research engine:

Scoper -> AAR Planner -> Fan-out -> Synthesis

Intelligence Research OS adds the repeatable intelligence layer:

Mission -> Source Packs -> Signals -> Deltas -> Evidence -> Synthesis -> Output Pack

Boundary

  • ResearchModule: core Python package and engine.
  • Intelligence Research OS: mission/use-case layer, UX, source packs, signal scoring, delta logic, opportunity briefs, and workflow-callable outputs.
  • Scheduling: external orchestration only. Cron, Codex automations, GitHub Actions, Airflow, Temporal, n8n, Zapier, or app schedulers invoke the engine.

Product Thesis

Competitive intelligence, social listening, content research, trend monitoring, and market intelligence exist as separate product categories. The opportunity is not to reinvent each one. The opportunity is to compose them into a tool chest where the whole is stronger than the parts:

  • source-backed research
  • strict provenance
  • multi-POV synthesis
  • delta detection
  • weak-signal interpretation
  • opportunity mining
  • role-specific outputs
  • callable workflows

Intelligence Layer Concepts

Concept Purpose
MissionSpec Defines the repeatable job to be done and the desired output.
WatchSpec Defines entities/topics, filters, exclusions, thresholds, and prior-run comparison rules.
SourcePack Defines source families such as Reddit, HN, GitHub, YouTube, X/Twitter, Substack, news, blogs, competitor sites, changelogs, reviews, and job posts.
Signal A classified observed item or event.
SignalCluster Related signals grouped into an emerging pattern.
DeltaReport What changed since the previous run.
OpportunityBrief Interprets signals into business, content, product, sales, or strategy opportunities.
OutputPack Role-specific deliverable: memo, battlecard, weekly pulse, content backlog, account brief, etc.

Signal Types

  • launch
  • pricing change
  • positioning change
  • customer complaint
  • new repo/tool
  • analyst mention
  • hiring signal
  • funding/acquisition
  • content trend
  • community debate
  • weak signal/anomaly
  • regulatory/policy shift

Delta Logic

Every recurring mission should distinguish:

  • new
  • recurring
  • accelerating
  • fading
  • contradictory
  • stale
  • anomalous

Output Packs

  • executive memo
  • sales battlecard
  • product opportunity brief
  • content idea backlog
  • weekly market pulse
  • competitor digest
  • account prep brief
  • investment/thesis memo
  • customer voice report
  • tool radar report

Non-Goals

  • Do not build a scheduler inside the engine.
  • Do not create a second provenance system outside ResearchModule artifacts.
  • Do not accept unsupported claims in intelligence outputs.
  • Do not make monitoring a free-form agent behavior; missions must be explicit and file-gated.

Current Snapshot

See wiki/current-state.md for the latest built/pending/next-step status. As of 2026-05-22, the core engine and local workbench spine are in place, the Competitive Intelligence demo workflow passes browser E2E, and the next unlock is wiring UI mission execution into the live file-gated engine pipeline.