Skip to content

INTEROP.md

Latest commit

 

History

History
80 lines (65 loc) · 4.08 KB

INTEROP.md

File metadata and controls

80 lines (65 loc) · 4.08 KB
Raw

Cognis interop map

How quotecraft fits the wider Cognis suite — a composable set of defensive, analytical tools that all run on your own hardware and speak plain JSON (or an OpenAI-compatible /v1). quotecraft lives in the General OSINT / security tool cluster.

graph LR
  subgraph infra["AI backbone (private)"]
    EM[edgemesh]:::hub
    MR[modelroute]; UF[uncensored-fleet]; CC[cognis-code]
    MR --> EM; UF --> EM; CC --> EM
  end
  subgraph cog["Agents + cognition"]
    AL[agentlex]; HM[humind]; AS[agentsmith]; EN[engram]
    HM -- speaks --> AL; AL -- facts --> HM; AS -- orchestrates --> AL; EN --- HM
  end
  subgraph dom["Domain intelligence"]
    MI[maritimeint]; GEO[geoint]; FIN[ownership/finance]; TI[threat-intel]; DR[drone/C-UAS]
  end
  EM -- "/v1 add-ins" --> dom
  EM -- "/v1 enrichment" --> cog
  dom -- "findings as messages" --> cog
  classDef hub fill:#6b46c1,color:#fff;
Loading

Key edges

from relation to
quotecraft JSON findings compose with awesome-cognis
quotecraft AI add-ins are served /v1 by edgemesh (your fleet)
quotecraft findings can be narrated / reasoned over by humind -> agentlex
quotecraft export to intel formats via stixgen (STIX) / attackmap (ATT&CK)

Composition patterns

Everything reads/writes JSON, so tools chain with ordinary pipes; nothing leaves the box.

1 — chain within the general osint / security tool cluster. quotecraft output feeds the next tool:

quotecraft ... --format json > out.json          # this tool's findings (see `quotecraft --help`)
awesome-cognis ... < out.json                      # the cluster sibling consumes them

2 — enrich with the private-AI backbone. Point add-ins at one /v1 for the whole fleet:

export OPENAI_BASE_URL=http://localhost:8080/v1   # an edgemesh gateway over your fleet
quotecraft ...                                         # vision / reasoning add-ins light up

3 — export findings to your SOC's formats.

quotecraft ... --format json | stixgen from-json > bundle.stix.json   # STIX 2.1
quotecraft ... --format json | attackmap map > attack.json            # ATT&CK techniques

4 — narrate through cognition + agents. humind extracts salience from quotecraft's output; agentlex holds it as KB facts and fires Horn rules to escalate. quotecraft slots into the Threat-intel export stack below.

Reference stacks

Pick the smallest stack that answers your question; each is a subset of the 300+ suite.

Stack Representative repos Flow
Sanctions-evasion screening maritimeint + OFAC/OFSI/EU/OpenSanctions feeds importers -> screen -> pass/fail gate
GEOINT fusion locateanything - geolens - geoaoi-pro imagery -> geolocate -> plot on AOI / MIL-STD-2525
Ownership & finance corpmap - cryptotrace - personagraph entity -> beneficial owner -> wallets -> identity
Threat-intel export stixgen - iocextract - attackmap - ttphunt findings -> STIX / IOCs -> ATT&CK -> hunt
Counter-UAS maritime maritimeint - uaslog - awesome-drone-warfare-osint AIS + drone telemetry -> correlated picture
Cognition + agents humind - agentlex - agentsmith - engram extract -> KB facts/rules -> orchestrate -> memory
Private-AI backbone edgemesh - modelroute - uncensored-fleet - cognis-code one /v1 powering every tool's add-ins

Every domain stack sits on top of the private-AI backbone: point any tool's add-ins at an edgemesh gateway (OPENAI_BASE_URL) and one fleet serves vision, reasoning, and narration to all of them.

Part of a cross-repo interop pass. 300+ tools -> github.com/cognis-digital - awesome-cognis