AntiFl0ck eagle mark

AntiFl0ck

Open-source counter-surveillance for the networks you control.

Map your exposure. Verify your route. Gate sensitive actions — with every decision signed into an audit log that never leaves your machine.

pre-alpha Apache-2.0 self-hosted evidence-first Go + React

The scenario

When the trusted path disappears, the action waits.

A laptop joins an untrusted network. An application asks to perform a sensitive upload. The trusted route is gone — so the action is held, with a recorded reason. When the route returns and is verified, the action is allowed. Five signed events, replayable on your machine today.

Demo lifecycle: trusted, route lost, action held, route verified, action allowed

The full coffee-shop scenario runs locally and is entirely simulated — no VPN account, no real data, no cloud dependency.

What it does

Four verbs

See

Map your devices, routes, mesh peers, and the metadata an outside observer could plausibly collect. An inventory of your own footprint — not a threat feed.

Understand

Label every finding by evidence status: detected, verified, reported, inferred, suspected, unknown. A guess never quietly becomes an accusation.

Decide

Deterministic, operator-defined policy allows, holds, or blocks sensitive actions. Same inputs, same decision, every time. No model guessing at risk.

Prove

Signed, hash-chained decisions with plain-language explanations, expirations, and rollback results — stored locally, never in a vendor's cloud.

Evidence model

A guess never quietly becomes an accusation.

Something directly observed is not the same as something a third party reported, and neither is the same as an inference. AntiFl0ck keeps those categories separate on purpose. An AI may explain a finding; it never makes the allow-or-block decision.

Evidence scale from DETECTED through UNKNOWN

Quickstart

Run the demo

Needs Docker and Node 24+. Nothing else — no account, no real data.

# clone and start the stack
git clone https://github.com/AetherAI3/AntiFlock
cd AntiFlock
make dev

# run the simulated coffee-shop scenario
make lab

# Third-Eye dashboard
http://127.0.0.1:4173  — user: operator

On Windows without make: npm run dev / npm run lab. Token is in .antiflock/dev.env. Full guide in the operator runbook.

Honest status

What works, and what doesn't

The local simulated slice is complete and sits behind a ten-gate release check. The real-world pieces are later milestones — they are not claimed until proven.

Working today

  • Coffee-shop simulation end-to-end, five signed audit events
  • Signed, hash-chained event and audit log (SQLite)
  • Deterministic policy engine and signed expiring plans
  • Secure-action gate + TypeScript SDK
  • Endpoint enrollment and mTLS device identity
  • Linux route and interface observation
  • Read-only Tailscale and Headscale mesh probes
  • Third-Eye dashboard

Open engineering work

  • Production enforcement and real packet-path integration
  • Mobile (Android) enforcement beyond the reference state machine
  • macOS and Windows collectors
  • Provider lifecycle automation — key rotation, revocation
  • Independent security and privacy review

Hard boundaries: release status · threat model · open decisions

Contribute

Five lanes in

Help build the defensive layer surveillance technology never gave ordinary people. Scoped issues are labelled good first issue and help wanted.

Observers

Linux, macOS, Windows, and Android collectors; provider adapters.

Visualizers

Topology, exposure maps, and audit timelines in React.

Policy builders

Deterministic rules and secure-action integrations.

Researchers

Metadata models, ALPR research, threat modeling.

Hardening

Signing, storage, privilege separation, testing.

Read this first

Pre-release defensive software. AntiFl0ck does not provide anonymity, does not prove that surveillance is occurring, and does not replace a VPN. It does not detect or locate ALPR cameras, scan for nearby hardware, or interfere with anyone's surveillance infrastructure — it is a defensive layer for the networks you already control.

AntiFl0ck is an independent open-source project. It is not affiliated with or endorsed by Flock Safety. Report vulnerabilities privately via SECURITY.md, never a public issue.