sif

install · usage · modules · docs · contribute

fast, concurrent recon to exploitation in one binary. every scanner shares one connection-pooled http client.

what is sif?

sif is a recon and exploitation scanner that runs the whole chain in one binary: subdomain enum, port scan, crawler, nuclei, framework/cve detection, js secret extraction, web-vuln probes (cors/xss/redirect), cloud and takeover checks. 25+ scan types, one command.

sif -u https://example.com -dnslist -ports -crawl -js -framework -nuclei

nuclei and colly are compiled in as libraries rather than shelled out to (there's no exec.Command in the tree), so it's a single static binary with no runtime dependencies and nothing to wire together.

every scanner runs through one shared http client and a work-stealing worker pool. -proxy, -H, -cookie and -rate-limit apply to the whole run at once, connections get pooled and reused across the scan (a single-host run reuses one connection for ~50 requests instead of dialing 50 times), and a slow host doesn't hold the rest up. that shared client is the practical reason to use it over piping a stack of separate tools together. port scanning is connect()-based, so rustscan and nmap are still faster at raw port scans.

it reads targets from stdin and prints findings one per line under -silent, so it composes:

subfinder -d example.com | sif -silent -crawl -js -nuclei | notify

-diff turns a re-scan into a monitor that only reports what changed, -notify posts to slack/discord/telegram/webhook, and runs export to sarif and markdown.

install

homebrew (macos)

brew tap vmfunc/sif
brew install sif

arch linux (aur)

install using your preferred aur helper:

yay -S sif
# or
paru -S sif

nix

# nixpkgs (declarative: add to configuration.nix or home-manager)
environment.systemPackages = [ pkgs.sif ];

# or imperatively
nix profile install nixpkgs#sif

# or just run it without installing
nix run nixpkgs#sif -- -u https://example.com -headers -sh -framework

the repo also ships a flake if you want to build from source:

nix run github:vmfunc/sif

debian/ubuntu (apt)

curl -1sLf 'https://dl.cloudsmith.io/public/sif/deb/setup.deb.sh' | sudo -E bash
sudo apt-get install sif

from releases

grab the latest binary from releases.

from source

git clone https://github.com/vmfunc/sif.git
cd sif
make

requires go 1.25+

aur (manual install)

git clone https://aur.archlinux.org/sif.git
cd sif
makepkg -si

usage

# basic scan
./sif -u https://example.com

# directory fuzzing
./sif -u https://example.com -dirlist medium

# subdomain enumeration
./sif -u https://example.com -dnslist medium

# port scanning
./sif -u https://example.com -ports common

# javascript framework detection + cloud misconfig
./sif -u https://example.com -js -c3

# shodan host intelligence (requires SHODAN_API_KEY env var)
./sif -u https://example.com -shodan

# securitytrails domain discovery (requires SECURITYTRAILS_API_KEY env var)
# discovers subdomains + associated domains, then scans all of them
./sif -u https://example.com -securitytrails -headers

# sql recon + lfi scanning
./sif -u https://example.com -sql -lfi

# web vuln probes (cors, open redirect, reflected xss)
./sif -u https://example.com -cors -redirect -xss

# framework detection (with cve lookup)
./sif -u https://example.com -framework

# a broad sweep
./sif -u https://example.com -dirlist small -dnslist small -ports common -headers -sh -cms -framework -git -whois

run ./sif -h for all options.

commands

a couple of subcommands run without scanning:

# print the version (release builds are stamped; local builds use git describe)
./sif version

# show the latest release notes (also -pn)
./sif patchnote

the first time you run a new release, sif prints that release's notes once. set SIF_NO_PATCHNOTES=1 to turn that off.

modules

sif has a modular architecture. modules are defined in yaml and can be extended by users.

built-in scan flags

http options

these apply to every outbound request across all scanners:

# scan through a socks5 proxy with a custom header, cookie and 20 req/s cap
./sif -u https://example.com -headers -proxy socks5://127.0.0.1:1080 -H "Authorization: Bearer tok" -cookie "session=abc" -rate-limit 20

a scanner that sets a header explicitly (e.g. an api key) always wins over the global default.

report export

write the run's findings out to a file for ci/cd or triage:

# scan and emit both a sarif and markdown report
./sif -u https://example.com -headers -cors -sarif out.sarif -md out.md

sarif output is ingestable by github code scanning; markdown is a readable per-target summary.

diff mode

-diff turns a re-scan into a monitor: sif snapshots each target's normalized findings to a json file, and on the next run reports only the delta (+ new / - gone) against that snapshot, then overwrites it. the first run for a target has no baseline, so everything is + new. snapshots land in -store (one sanitized file per target); when unset they reuse the log dir, falling back to <user-config>/sif/state.

# baseline run, then re-scan later and see only what moved
./sif -u https://example.com -sh -cors -diff
./sif -u https://example.com -sh -cors -diff

the snapshot is always rewritten, so each run diffs against the previous one. the delta is chrome (it rides the normal output sink / stderr under -silent), not the findings stream.

notify

ship findings to a chat/webhook sink so a continuous-recon run alerts on what it turns up. every provider is a single POST through the shared http client, so the global proxy/rate-limit/header config applies.

providers are configured env-first; a yaml file (-notify-config) overrides per-field. the yaml keys match projectdiscovery/notify so an existing config ports over:

# alert slack on medium+ findings discovered during a scan
export SLACK_WEBHOOK_URL=https://hooks.slack.com/services/...
./sif -u https://example.com -cors -xss -notify -notify-severity medium

a provider with no destination is skipped; with nothing configured, -notify is a silent no-op. slack/discord/telegram receive a fixed-width finding block; the generic webhook receives structured json ({count, findings[]}).

pipe mode

sif reads targets from stdin and accepts naked hosts, so it drops into a unix pipeline. -silent routes all banner/spinner/log chrome to stderr and prints one normalized finding per line ([severity] target module title) to stdout:

# subfinder feeds hosts, sif probes them, notify ships the findings
subfinder -d example.com | sif -silent -probe | notify

scheme-less hosts default to https://; an explicit http:///https:// is kept; any other scheme (ftp://, ...) is rejected.

yaml modules

list available modules:

./sif -lm

run specific modules:

# run by id
./sif -u https://example.com -m sqli-error-based,xss-reflected

# run by tag
./sif -u https://example.com -mt owasp-top10

# run all modules
./sif -u https://example.com -am

custom modules

create your own modules in ~/.config/sif/modules/. modules use a yaml format similar to nuclei templates:

id: my-custom-check
info:
  name: my custom security check
  author: you
  severity: medium
  description: checks for something specific
  tags: [custom, recon]

type: http

http:
  method: GET
  paths:
    - "{{BaseURL}}/admin"
    - "{{BaseURL}}/login"

  matchers:
    - type: status
      status:
        - 200

    - type: word
      part: body
      words:
        - "admin panel"
        - "login"
      condition: or

see docs/modules.md for the full module format.

contribute

contributions welcome. see contributing.md for guidelines.

# format
gofmt -w .

# lint
go run github.com/golangci/golangci-lint/v2/cmd/[email protected] run

# test
go test ./...

community

join our discord for support, feature discussions, and pentesting tips:

contributors

acknowledgements

• projectdiscovery for nuclei and other security tools
• shodan for infrastructure intelligence