agentleFS
Sign inStart free
ThirdKeyAIThirdKeyAI/agentsniff/CLAUDE.md

agentsniff

CLAUDE.md15 starsChanged 7 days agoApache-2.0View on GitHub

AgentSniff is a Rust workspace for authorized AI-agent network discovery. Raw capture requires appropriate Linux capabilities. Keep privileged Docker and host-network use opt-in.

  • In this repositoryOne other file it gives its agents.

The file

Create a free account to mount
Files
CLAUDE.md

AgentSniff Development Guide

AgentSniff is a Rust workspace for authorized AI-agent network discovery.

Stack

  • Rust 2021 workspace with Tokio
  • Axum dashboard/API
  • Reqwest active probes
  • SQLite persistence through rusqlite
  • Single-file embedded HTML/JS/CSS dashboard
  • Optional eBPF, Zeek, service enrichment (naabu/httpx/zgrab2), and Nmap XML import

Layout

Cargo.toml
crates/agentsniff/
  src/
    main.rs
    scanner.rs
    server.rs
    config.rs
    models.rs
    fusion.rs
    storage*.rs
    notifier.rs
    gate/            # forward-auth server; engine lives in agentsniff-gate
    detectors/
    integrations/
    signatures/      # dataset updater; data + verifier live in agentsniff-signatures
  assets/
    index.html
  tests/
crates/agentsniff-signatures/
  src/lib.rs
  assets/signatures/   # signed datasets + signing_key.pub
crates/agentsniff-gate/
crates/agentsniff-common/
scripts/
docs/

Invariants

  • AgentPin shape validation is discovery only. Do not use “verified” or CONFIRMED without cryptographic identity validation.
  • Detection datasets must verify against crates/agentsniff-signatures/assets/signatures/signing_key.pub; never trust a key supplied by a sidecar.
  • Signature updates are complete-set, fail-closed, and atomically activated.
  • The dashboard binds to loopback by default. A non-loopback listener requires a 32-byte-or-longer API token, and every non-health route must remain authenticated when a token is configured.
  • SSE is read-only; scan creation is a JSON POST.
  • Network-derived strings must be inert in HTML and terminal output.
  • Enforce target and response-size budgets before allocation.
  • The gate never fetches issuer-controlled URLs (Web Bot Auth key directories); keys are operator-pinned. Forwarded headers count only from trusted proxies. The gate defaults to monitor mode and loopback.

Commands

cargo fmt --all -- --check
cargo test --workspace
cargo clippy --workspace --lib --bins -- -D warnings
cargo run -p agentsniff -- scan 192.168.1.0/24
cargo run -p agentsniff -- serve --port 9090

Raw capture requires appropriate Linux capabilities. Keep privileged Docker and host-network use opt-in.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free