SuperAPI AI Directory

osint-recon-automation

General & other

View repository

Updated 10-16 06:45

About this skill

OSINT Recon Automation

Authorization Boundary

  • Restrict to your own assets, bug bounty programs with explicit scope, or assignments with written authorization.
  • No social engineering, doxing, harassment, or targeting individuals.
  • Respect rate limits and provider terms; never weaponize discovered credentials.

Recon Workflow

  1. Seed: registered domains, ASNs, org names, GitHub orgs, app store handles.
  2. Expand passively: cert transparency, passive DNS, reverse WHOIS, ASN ranges, code search, leaked credential repos, archive snapshots.
  3. Resolve and validate: dnsx, httpx, screenshot with gowitness or aquatone.
  4. Pivot: shared favicons (favfreak), Google Analytics IDs, S3 bucket naming, JS endpoints, ASN neighbors.
  5. Dedupe and graph: build a node-edge model (asset)-[relation]->(asset) and store in Neo4j or JSONL.
  6. Monitor: diff snapshots; alert on new subdomain, new exposed service, leaked secret, or impersonation domain.

Exposure Signals

  • Exposed admin panels, dev/stage hosts, open S3/Blob/GCS, leaked .env, .git/, .DS_Store, swagger, GraphQL introspection, package registry takeover candidates, dangling DNS.

Output Contract

  • seeds.yaml, assets.jsonl, graph.json, screenshots/, exposures.csv, diff/<date>.md.
  • report.md: new exposures since last run with severity and owner.
  • takedowns.md: impersonation domains and reporting paths.