Compare commits

...

15 Commits

Author SHA1 Message Date
Xe Iaso
0b2e2c34df Revert "Revert "lib/anubis: support setting extended cookie flags (#120)" (#134)"
This reverts commit ecc6b47f90.
2025-03-26 20:50:58 -04:00
Xe Iaso
ecc6b47f90 Revert "lib/anubis: support setting extended cookie flags (#120)" (#134)
This reverts commit e7cbd349f3.
2025-03-26 20:50:54 -04:00
Xe Iaso
e7cbd349f3 lib/anubis: support setting extended cookie flags (#120)
* lib/anubis: support setting extended cookie flags

Signed-off-by: Xe Iaso <me@xeiaso.net>

* lib: use cookie name consistently

Signed-off-by: Xe Iaso <me@xeiaso.net>

---------

Signed-off-by: Xe Iaso <me@xeiaso.net>
2025-03-26 19:04:18 -04:00
Henri Vasserman
07bb5f63f9 fix(docs): Make dark mode diff lines readable (#130)
* fix(docs): Make dark mode diff lines readable

If using dark mode, these lines are not legible at all. I separated the colors into variables and added
more contrasting colors for the dark mode.

* chore: add to changelog
2025-03-26 15:44:20 -04:00
Xe Iaso
4155719422 cmd/anubis: allow setting key bytes in flag/envvar (#97)
* cmd/anubis: allow setting key bytes in flag/envvar

Docs are updated to generate a random key on load and when people press
the recycle button.

Signed-off-by: Xe Iaso <me@xeiaso.net>

* review feedback fixups

Signed-off-by: Xe Iaso <me@xeiaso.net>

* Update cmd/anubis/main.go

Signed-off-by: Xe Iaso <me@xeiaso.net>

* Apply suggestions from code review

Co-authored-by: Ryan Cao <70191398+ryanccn@users.noreply.github.com>
Signed-off-by: Xe Iaso <me@xeiaso.net>

---------

Signed-off-by: Xe Iaso <me@xeiaso.net>
Co-authored-by: Ryan Cao <70191398+ryanccn@users.noreply.github.com>
2025-03-25 17:02:48 -04:00
Yulian Kuncheff
f29a200f09 Linting and staticcheck fixes. (#101)
* Fix linting and staticcheck issues

* Add changelog update

* Remove SetNext
2025-03-25 10:02:05 -04:00
soopyc
18cd8a66a2 docs: minor updates (#98)
* use _ instead of * for italicized text by convention

* remove mention of the `anubis` tag from /x/
2025-03-23 23:49:12 -04:00
Xe Iaso
725e11d3a6 lib: fix default difficulty (#96)
Before this did not respect the difficulty flag and instead used
difficulty 4. This has been fixed.

Signed-off-by: Xe Iaso <me@xeiaso.net>
2025-03-23 18:46:01 -04:00
Jared Allard
f462209b02 chore: remove built binary and prevent accidental addition again (#91)
Removes `main` from the repo and updates `.gitignore` to help prevent
accidents like this again. `.test` comes from go test binaries, which I
feel is also worth excluding

Signed-off-by: Jared Allard <jaredallard@users.noreply.github.com>
2025-03-22 22:44:11 -04:00
Jared Allard
acf5586e83 docs(README): fix mascot link (#88)
Title says it does, noticed it broke 😢

Signed-off-by: Jared Allard <jaredallard@users.noreply.github.com>
2025-03-22 22:37:16 -04:00
dependabot[bot]
9d68e73d03 build(deps): bump github.com/go-jose/go-jose/v3 from 3.0.3 to 3.0.4 (#89)
Bumps [github.com/go-jose/go-jose/v3](https://github.com/go-jose/go-jose) from 3.0.3 to 3.0.4.
- [Release notes](https://github.com/go-jose/go-jose/releases)
- [Changelog](https://github.com/go-jose/go-jose/blob/main/CHANGELOG.md)
- [Commits](https://github.com/go-jose/go-jose/compare/v3.0.3...v3.0.4)

---
updated-dependencies:
- dependency-name: github.com/go-jose/go-jose/v3
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2025-03-22 21:18:02 -04:00
Yulian Kuncheff
6156d3d729 Refactor and split out things into cmd and lib (#77)
* Refactor anubis to split business logic into a lib, and cmd to just be direct usage.

* Post-rebase fixes.

* Update changelog, remove unnecessary one.

* lib: refactor this

This is mostly based on my personal preferences for how Go code should
be laid out. I'm not sold on the package name "lib" (I'd call it anubis
but that would stutter), but people are probably gonna import it as
libanubis so it's likely fine.

Packages have been "flattened" to centralize implementation with area of
concern. This goes against the Java-esque style that many people like,
but I think this helps make things simple.

Most notably: the dnsbl client (which is a hack) is an internal package
until it's made more generic. Then it can be made external.

I also fixed the logic such that `go generate` works and rebased on
main.

* internal/test: run tests iff npx exists and DONT_USE_NETWORK is not set

Signed-off-by: Xe Iaso <me@xeiaso.net>

* internal/test: install deps

Signed-off-by: Xe Iaso <me@xeiaso.net>

* .github/workflows: verbose go tests?

Signed-off-by: Xe Iaso <me@xeiaso.net>

* internal/test: sleep 2

Signed-off-by: Xe Iaso <me@xeiaso.net>

* internal/test: nix this test so CI works

Signed-off-by: Xe Iaso <me@xeiaso.net>

* internal/test: warmup per browser?

Signed-off-by: Xe Iaso <me@xeiaso.net>

* internal/test: disable for now :(

Signed-off-by: Xe Iaso <me@xeiaso.net>

* lib/anubis: do not apply bot rules if address check fails

Closes #83

---------

Signed-off-by: Xe Iaso <me@xeiaso.net>
Co-authored-by: Xe Iaso <me@xeiaso.net>
2025-03-22 18:44:49 -04:00
Valentin Anger
af6f05554f internal/test: introduce integration tests using Playwright (#81) 2025-03-22 16:36:27 -04:00
Dennis ten Hoove
1509b06cb9 Cleanup regex (#66)
* Cleanup regex

Were were going overkill on the escape characters

* Update docs/docs/CHANGELOG.md

Co-authored-by: Xe Iaso <me@xeiaso.net>
Signed-off-by: Dennis ten Hoove <36002865+dennis1248@users.noreply.github.com>

---------

Signed-off-by: Dennis ten Hoove <36002865+dennis1248@users.noreply.github.com>
Co-authored-by: Xe Iaso <me@xeiaso.net>
2025-03-22 13:52:14 -04:00
Xe Iaso
56cdb2e51b Fix docker image CI for pull requests (#84)
Closes #65

Pull request images will now be `ttl.sh/techaro/pr-{number}/anubis:24h`.
2025-03-22 11:26:49 -04:00
63 changed files with 1750 additions and 860 deletions

44
.github/workflows/docker-pr.yml vendored Normal file
View File

@@ -0,0 +1,44 @@
name: Docker image builds (pull requests)
on:
pull_request:
branches: [ "main" ]
env:
DOCKER_METADATA_SET_OUTPUT_ENV: "true"
permissions:
contents: read
jobs:
build:
runs-on: ubuntu-latest
steps:
- name: Checkout code
uses: actions/checkout@v4
with:
fetch-tags: true
fetch-depth: 0
- uses: actions/setup-go@v5
with:
go-version: '1.24.x'
- uses: ko-build/setup-ko@v0.8
- name: Docker meta
id: meta
uses: docker/metadata-action@v5
with:
images: ghcr.io/techarohq/anubis
- name: Build and push
id: build
run: |
go run ./cmd/containerbuild --docker-repo ghcr.io/techarohq/anubis --slog-level debug
env:
PULL_REQUEST_ID: ${{ github.event.number }}
- run: |
echo "Test this with:"
echo "docker pull ${{ steps.build.outputs.docker_image }}"

View File

@@ -5,8 +5,6 @@ on:
push:
branches: [ "main" ]
tags: [ "v*" ]
pull_request:
branches: [ "main" ]
env:
DOCKER_METADATA_SET_OUTPUT_ENV: "true"
@@ -28,12 +26,6 @@ jobs:
fetch-tags: true
fetch-depth: 0
- name: Set up QEMU
uses: docker/setup-qemu-action@v3
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
- uses: actions/setup-go@v5
with:
go-version: '1.24.x'
@@ -57,20 +49,6 @@ jobs:
id: build
run: |
go run ./cmd/containerbuild --docker-repo ghcr.io/techarohq/anubis --slog-level debug
env:
PULL_REQUEST_ID: ${{ github.event.number }}
# - name: "Comment about where to test this"
# uses: thollander/actions-comment-pull-request@v3
# if: ${{github.event_name == 'pull_request'}}
# with:
# message: |
# You can try this PR out by using the following docker image:
# ```
# ${{ steps.build.outputs.docker_image }}
# ```
# comment-tag: ${{ steps.build.outputs.docker_image }}
- name: Generate artifact attestation
uses: actions/attest-build-provenance@v2

View File

@@ -56,6 +56,14 @@ jobs:
restore-keys: |
${{ runner.os }}-golang-
- name: Cache playwright binaries
uses: actions/cache@v3
id: playwright-cache
with:
path: |
~/.cache/ms-playwright
key: ${{ runner.os }}-playwright-${{ hashFiles('**/go.sum') }}
- name: Build
run: go build ./...

6
.gitignore vendored
View File

@@ -1,2 +1,6 @@
.env
*.rpm
*.rpm
# Go binaries and test artifacts
main
*.test

View File

@@ -1,7 +1,7 @@
# Anubis
<center>
<img width=256 src="./cmd/anubis/static/img/happy.webp" alt="A smiling chibi dark-skinned anthro jackal with brown hair and tall ears looking victorious with a thumbs-up" />
<img width=256 src="./web/static/img/happy.webp" alt="A smiling chibi dark-skinned anthro jackal with brown hair and tall ears looking victorious with a thumbs-up" />
</center>
![enbyware](https://pride-badges.pony.workers.dev/static/v1?label=enbyware&labelColor=%23555&stripeWidth=8&stripeColors=FCF434%2CFFFFFF%2C9C59D1%2C2C2C2C)

19
anubis.go Normal file
View File

@@ -0,0 +1,19 @@
// Package Anubis contains the version number of Anubis.
package anubis
// Version is the current version of Anubis.
//
// This variable is set at build time using the -X linker flag. If not set,
// it defaults to "devel".
var Version = "devel"
// CookieName is the name of the cookie that Anubis uses in order to validate
// access.
const CookieName = "within.website-x-cmd-anubis-auth"
// StaticPath is the location where all static Anubis assets are located.
const StaticPath = "/.within.website/x/cmd/anubis/"
// DefaultDifficulty is the default "difficulty" (number of leading zeroes)
// that must be met by the client in order to pass the challenge.
const DefaultDifficulty = 4

View File

@@ -1,5 +0,0 @@
# CHANGELOG
## 2025-01-24
- Added support for custom bot policy documentation, allowing administrators to change how Anubis works to meet their needs.

View File

@@ -4,18 +4,11 @@ import (
"context"
"crypto/ed25519"
"crypto/rand"
"crypto/sha256"
"crypto/subtle"
"embed"
"encoding/hex"
"encoding/json"
"flag"
"fmt"
"io"
"log"
"log/slog"
"math"
mrand "math/rand"
"net"
"net/http"
"net/http/httputil"
@@ -29,73 +22,45 @@ import (
"time"
"github.com/TecharoHQ/anubis"
"github.com/TecharoHQ/anubis/cmd/anubis/internal/config"
"github.com/TecharoHQ/anubis/cmd/anubis/internal/dnsbl"
"github.com/TecharoHQ/anubis/internal"
"github.com/TecharoHQ/anubis/xess"
"github.com/a-h/templ"
libanubis "github.com/TecharoHQ/anubis/lib"
"github.com/TecharoHQ/anubis/lib/policy/config"
"github.com/TecharoHQ/anubis/web"
"github.com/facebookgo/flagenv"
"github.com/golang-jwt/jwt/v5"
"github.com/prometheus/client_golang/prometheus"
"github.com/prometheus/client_golang/prometheus/promauto"
"github.com/prometheus/client_golang/prometheus/promhttp"
)
var (
bind = flag.String("bind", ":8923", "network address to bind HTTP to")
bindNetwork = flag.String("bind-network", "tcp", "network family to bind HTTP to, e.g. unix, tcp")
challengeDifficulty = flag.Int("difficulty", defaultDifficulty, "difficulty of the challenge")
metricsBind = flag.String("metrics-bind", ":9090", "network address to bind metrics to")
metricsBindNetwork = flag.String("metrics-bind-network", "tcp", "network family for the metrics server to bind to")
socketMode = flag.String("socket-mode", "0770", "socket mode (permissions) for unix domain sockets.")
robotsTxt = flag.Bool("serve-robots-txt", false, "serve a robots.txt file that disallows all robots")
policyFname = flag.String("policy-fname", "", "full path to anubis policy document (defaults to a sensible built-in policy)")
slogLevel = flag.String("slog-level", "INFO", "logging level (see https://pkg.go.dev/log/slog#hdr-Levels)")
target = flag.String("target", "http://localhost:3923", "target to reverse proxy to")
healthcheck = flag.Bool("healthcheck", false, "run a health check against Anubis")
debugXRealIPDefault = flag.String("debug-x-real-ip-default", "", "If set, replace empty X-Real-Ip headers with this value, useful only for debugging Anubis and running it locally")
//go:embed static botPolicies.json
static embed.FS
challengesIssued = promauto.NewCounter(prometheus.CounterOpts{
Name: "anubis_challenges_issued",
Help: "The total number of challenges issued",
})
challengesValidated = promauto.NewCounter(prometheus.CounterOpts{
Name: "anubis_challenges_validated",
Help: "The total number of challenges validated",
})
droneBLHits = promauto.NewCounterVec(prometheus.CounterOpts{
Name: "anubis_dronebl_hits",
Help: "The total number of hits from DroneBL",
}, []string{"status"})
failedValidations = promauto.NewCounter(prometheus.CounterOpts{
Name: "anubis_failed_validations",
Help: "The total number of failed validations",
})
timeTaken = promauto.NewHistogram(prometheus.HistogramOpts{
Name: "anubis_time_taken",
Help: "The time taken for a browser to generate a response (milliseconds)",
Buckets: prometheus.ExponentialBucketsRange(1, math.Pow(2, 18), 19),
})
bind = flag.String("bind", ":8923", "network address to bind HTTP to")
bindNetwork = flag.String("bind-network", "tcp", "network family to bind HTTP to, e.g. unix, tcp")
cookieDomain = flag.String("cookie-domain", "", "if set, the top-level domain that the Anubis cookie will be valid for")
cookieName = flag.String("cookie-name", anubis.CookieName, "the name of the cookie that Anubis stores challenge pass records in")
cookiePartitioned = flag.Bool("cookie-partitioned", false, "if true, sets the partitioned flag on Anubis cookies, enabling CHIPS support")
challengeDifficulty = flag.Int("difficulty", anubis.DefaultDifficulty, "difficulty of the challenge")
ed25519PrivateKeyHex = flag.String("ed25519-private-key-hex", "", "private key used to sign JWTs, if not set a random one will be assigned")
metricsBind = flag.String("metrics-bind", ":9090", "network address to bind metrics to")
metricsBindNetwork = flag.String("metrics-bind-network", "tcp", "network family for the metrics server to bind to")
socketMode = flag.String("socket-mode", "0770", "socket mode (permissions) for unix domain sockets.")
robotsTxt = flag.Bool("serve-robots-txt", false, "serve a robots.txt file that disallows all robots")
policyFname = flag.String("policy-fname", "", "full path to anubis policy document (defaults to a sensible built-in policy)")
slogLevel = flag.String("slog-level", "INFO", "logging level (see https://pkg.go.dev/log/slog#hdr-Levels)")
target = flag.String("target", "http://localhost:3923", "target to reverse proxy to")
healthcheck = flag.Bool("healthcheck", false, "run a health check against Anubis")
debugXRealIPDefault = flag.String("debug-x-real-ip-default", "", "If set, replace empty X-Real-Ip headers with this value, useful only for debugging Anubis and running it locally")
)
const (
cookieName = "within.website-x-cmd-anubis-auth"
staticPath = "/.within.website/x/cmd/anubis/"
defaultDifficulty = 4
)
func keyFromHex(value string) (ed25519.PrivateKey, error) {
keyBytes, err := hex.DecodeString(value)
if err != nil {
return nil, fmt.Errorf("supplied key is not hex-encoded: %w", err)
}
//go:generate go tool github.com/a-h/templ/cmd/templ generate
//go:generate esbuild js/main.mjs --sourcemap --bundle --minify --outfile=static/js/main.mjs
//go:generate gzip -f -k static/js/main.mjs
//go:generate zstd -f -k --ultra -22 static/js/main.mjs
//go:generate brotli -fZk static/js/main.mjs
if len(keyBytes) != ed25519.SeedSize {
return nil, fmt.Errorf("supplied key is not %d bytes long, got %d bytes", ed25519.SeedSize, len(keyBytes))
}
return ed25519.NewKeyFromSeed(keyBytes), nil
}
func doHealthCheck() error {
resp, err := http.Get("http://localhost" + *metricsBind + "/metrics")
@@ -145,6 +110,34 @@ func setupListener(network string, address string) (net.Listener, string) {
return listener, formattedAddress
}
func makeReverseProxy(target string) (http.Handler, error) {
u, err := url.Parse(target)
if err != nil {
return nil, fmt.Errorf("failed to parse target URL: %w", err)
}
transport := http.DefaultTransport.(*http.Transport).Clone()
// https://github.com/oauth2-proxy/oauth2-proxy/blob/4e2100a2879ef06aea1411790327019c1a09217c/pkg/upstream/http.go#L124
if u.Scheme == "unix" {
// clean path up so we don't use the socket path in proxied requests
addr := u.Path
u.Path = ""
// tell transport how to dial unix sockets
transport.DialContext = func(ctx context.Context, _, _ string) (net.Conn, error) {
dialer := net.Dialer{}
return dialer.DialContext(ctx, "unix", addr)
}
// tell transport how to handle the unix url scheme
transport.RegisterProtocol("unix", libanubis.UnixRoundTripper{Transport: transport})
}
rp := httputil.NewSingleHostReverseProxy(u)
rp.Transport = transport
return rp, nil
}
func main() {
flagenv.Parse()
flag.Parse()
@@ -158,13 +151,18 @@ func main() {
return
}
s, err := New(*target, *policyFname)
rp, err := makeReverseProxy(*target)
if err != nil {
log.Fatal(err)
log.Fatalf("can't make reverse proxy: %v", err)
}
policy, err := libanubis.LoadPoliciesOrDefault(*policyFname, *challengeDifficulty)
if err != nil {
log.Fatalf("can't parse policy file: %v", err)
}
fmt.Println("Rule error IDs:")
for _, rule := range s.policy.Bots {
for _, rule := range policy.Bots {
if rule.Action != config.RuleDeny {
continue
}
@@ -178,25 +176,29 @@ func main() {
}
fmt.Println()
mux := http.NewServeMux()
xess.Mount(mux)
var priv ed25519.PrivateKey
if *ed25519PrivateKeyHex == "" {
_, priv, err = ed25519.GenerateKey(rand.Reader)
if err != nil {
log.Fatalf("failed to generate ed25519 key: %v", err)
}
mux.Handle(staticPath, internal.UnchangingCache(http.StripPrefix(staticPath, http.FileServerFS(static))))
slog.Warn("generating random key, Anubis will have strange behavior when multiple instances are behind the same load balancer target, for more information: see https://anubis.techaro.lol/docs/admin/installation#key-generation")
} else {
priv, err = keyFromHex(*ed25519PrivateKeyHex)
if err != nil {
log.Fatalf("failed to parse and validate ED25519_PRIVATE_KEY_HEX: %v", err)
}
}
// mux.HandleFunc("GET /.within.website/x/cmd/anubis/static/js/main.mjs", serveMainJSWithBestEncoding)
mux.HandleFunc("POST /.within.website/x/cmd/anubis/api/make-challenge", s.makeChallenge)
mux.HandleFunc("GET /.within.website/x/cmd/anubis/api/pass-challenge", s.passChallenge)
mux.HandleFunc("GET /.within.website/x/cmd/anubis/api/test-error", s.testError)
if *robotsTxt {
mux.HandleFunc("/robots.txt", func(w http.ResponseWriter, r *http.Request) {
http.ServeFileFS(w, r, static, "static/robots.txt")
})
mux.HandleFunc("/.well-known/robots.txt", func(w http.ResponseWriter, r *http.Request) {
http.ServeFileFS(w, r, static, "static/robots.txt")
})
s, err := libanubis.New(libanubis.Options{
Next: rp,
Policy: policy,
ServeRobotsTXT: *robotsTxt,
PrivateKey: priv,
})
if err != nil {
log.Fatalf("can't construct libanubis.Server: %v", err)
}
wg := new(sync.WaitGroup)
@@ -209,10 +211,8 @@ func main() {
go metricsServer(ctx, wg.Done)
}
mux.HandleFunc("/", s.maybeReverseProxy)
var h http.Handler
h = mux
h = s
h = internal.DefaultXRealIP(*debugXRealIPDefault, h)
h = internal.XForwardedForToXRealIP(h)
@@ -267,428 +267,6 @@ func metricsServer(ctx context.Context, done func()) {
}
}
func sha256sum(text string) string {
hash := sha256.New()
hash.Write([]byte(text))
return hex.EncodeToString(hash.Sum(nil))
}
func (s *Server) challengeFor(r *http.Request, difficulty int) string {
fp := sha256.Sum256(s.priv.Seed())
data := fmt.Sprintf(
"Accept-Language=%s,X-Real-IP=%s,User-Agent=%s,WeekTime=%s,Fingerprint=%x,Difficulty=%d",
r.Header.Get("Accept-Language"),
r.Header.Get("X-Real-Ip"),
r.UserAgent(),
time.Now().UTC().Round(24*7*time.Hour).Format(time.RFC3339),
fp,
difficulty,
)
return sha256sum(data)
}
func New(target, policyFname string) (*Server, error) {
u, err := url.Parse(target)
if err != nil {
return nil, fmt.Errorf("failed to parse target URL: %w", err)
}
pub, priv, err := ed25519.GenerateKey(rand.Reader)
if err != nil {
return nil, fmt.Errorf("failed to generate ed25519 key: %w", err)
}
transport := http.DefaultTransport.(*http.Transport).Clone()
// https://github.com/oauth2-proxy/oauth2-proxy/blob/4e2100a2879ef06aea1411790327019c1a09217c/pkg/upstream/http.go#L124
if u.Scheme == "unix" {
// clean path up so we don't use the socket path in proxied requests
addr := u.Path
u.Path = ""
// tell transport how to dial unix sockets
transport.DialContext = func(ctx context.Context, _, _ string) (net.Conn, error) {
dialer := net.Dialer{}
return dialer.DialContext(ctx, "unix", addr)
}
// tell transport how to handle the unix url scheme
transport.RegisterProtocol("unix", unixRoundTripper{Transport: transport})
}
rp := httputil.NewSingleHostReverseProxy(u)
rp.Transport = transport
var fin io.ReadCloser
if policyFname != "" {
fin, err = os.Open(policyFname)
if err != nil {
return nil, fmt.Errorf("can't parse policy file %s: %w", policyFname, err)
}
} else {
policyFname = "(static)/botPolicies.json"
fin, err = static.Open("botPolicies.json")
if err != nil {
return nil, fmt.Errorf("[unexpected] can't parse builtin policy file %s: %w", policyFname, err)
}
}
defer fin.Close()
policy, err := parseConfig(fin, policyFname, *challengeDifficulty)
if err != nil {
return nil, err // parseConfig sets a fancy error for us
}
return &Server{
rp: rp,
priv: priv,
pub: pub,
policy: policy,
dnsblCache: NewDecayMap[string, dnsbl.DroneBLResponse](),
}, nil
}
// https://github.com/oauth2-proxy/oauth2-proxy/blob/master/pkg/upstream/http.go#L124
type unixRoundTripper struct {
Transport *http.Transport
}
// set bare minimum stuff
func (t unixRoundTripper) RoundTrip(req *http.Request) (*http.Response, error) {
req = req.Clone(req.Context())
if req.Host == "" {
req.Host = "localhost"
}
req.URL.Host = req.Host // proxy error: no Host in request URL
req.URL.Scheme = "http" // make http.Transport happy and avoid an infinite recursion
return t.Transport.RoundTrip(req)
}
type Server struct {
rp *httputil.ReverseProxy
priv ed25519.PrivateKey
pub ed25519.PublicKey
policy *ParsedConfig
dnsblCache *DecayMap[string, dnsbl.DroneBLResponse]
}
func (s *Server) maybeReverseProxy(w http.ResponseWriter, r *http.Request) {
lg := slog.With(
"user_agent", r.UserAgent(),
"accept_language", r.Header.Get("Accept-Language"),
"priority", r.Header.Get("Priority"),
"x-forwarded-for",
r.Header.Get("X-Forwarded-For"),
"x-real-ip", r.Header.Get("X-Real-Ip"),
)
cr, rule, err := s.check(r)
if err != nil {
lg.Error("check failed", "err", err)
templ.Handler(base("Oh noes!", errorPage("Internal Server Error: administrator has misconfigured Anubis. Please contact the administrator and ask them to look for the logs around \"maybeReverseProxy\"")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
r.Header.Add("X-Anubis-Rule", cr.Name)
r.Header.Add("X-Anubis-Action", string(cr.Rule))
lg = lg.With("check_result", cr)
policyApplications.WithLabelValues(cr.Name, string(cr.Rule)).Add(1)
ip := r.Header.Get("X-Real-Ip")
if s.policy.DNSBL && ip != "" {
resp, ok := s.dnsblCache.Get(ip)
if !ok {
lg.Debug("looking up ip in dnsbl")
resp, err := dnsbl.Lookup(ip)
if err != nil {
lg.Error("can't look up ip in dnsbl", "err", err)
}
s.dnsblCache.Set(ip, resp, 24*time.Hour)
droneBLHits.WithLabelValues(resp.String()).Inc()
}
if resp != dnsbl.AllGood {
lg.Info("DNSBL hit", "status", resp.String())
templ.Handler(base("Oh noes!", errorPage(fmt.Sprintf("DroneBL reported an entry: %s, see https://dronebl.org/lookup?ip=%s", resp.String(), ip))), templ.WithStatus(http.StatusOK)).ServeHTTP(w, r)
return
}
}
switch cr.Rule {
case config.RuleAllow:
lg.Debug("allowing traffic to origin (explicit)")
s.rp.ServeHTTP(w, r)
return
case config.RuleDeny:
clearCookie(w)
lg.Info("explicit deny")
if rule == nil {
lg.Error("rule is nil, cannot calculate checksum")
templ.Handler(base("Oh noes!", errorPage("Other internal server error (contact the admin)")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
hash, err := rule.Hash()
if err != nil {
lg.Error("can't calculate checksum of rule", "err", err)
templ.Handler(base("Oh noes!", errorPage("Other internal server error (contact the admin)")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
lg.Debug("rule hash", "hash", hash)
templ.Handler(base("Oh noes!", errorPage(fmt.Sprintf("Access Denied: error code %s", hash))), templ.WithStatus(http.StatusOK)).ServeHTTP(w, r)
return
case config.RuleChallenge:
lg.Debug("challenge requested")
default:
clearCookie(w)
templ.Handler(base("Oh noes!", errorPage("Other internal server error (contact the admin)")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
ckie, err := r.Cookie(cookieName)
if err != nil {
lg.Debug("cookie not found", "path", r.URL.Path)
clearCookie(w)
s.renderIndex(w, r)
return
}
if err := ckie.Valid(); err != nil {
lg.Debug("cookie is invalid", "err", err)
clearCookie(w)
s.renderIndex(w, r)
return
}
if time.Now().After(ckie.Expires) && !ckie.Expires.IsZero() {
lg.Debug("cookie expired", "path", r.URL.Path)
clearCookie(w)
s.renderIndex(w, r)
return
}
token, err := jwt.ParseWithClaims(ckie.Value, jwt.MapClaims{}, func(token *jwt.Token) (interface{}, error) {
return s.pub, nil
}, jwt.WithExpirationRequired(), jwt.WithStrictDecoding())
if err != nil || !token.Valid {
lg.Debug("invalid token", "path", r.URL.Path, "err", err)
clearCookie(w)
s.renderIndex(w, r)
return
}
if randomJitter() {
r.Header.Add("X-Anubis-Status", "PASS-BRIEF")
lg.Debug("cookie is not enrolled into secondary screening")
s.rp.ServeHTTP(w, r)
return
}
claims, ok := token.Claims.(jwt.MapClaims)
if !ok {
lg.Debug("invalid token claims type", "path", r.URL.Path)
clearCookie(w)
s.renderIndex(w, r)
return
}
challenge := s.challengeFor(r, rule.Challenge.Difficulty)
if claims["challenge"] != challenge {
lg.Debug("invalid challenge", "path", r.URL.Path)
clearCookie(w)
s.renderIndex(w, r)
return
}
var nonce int
if v, ok := claims["nonce"].(float64); ok {
nonce = int(v)
}
calcString := fmt.Sprintf("%s%d", challenge, nonce)
calculated := sha256sum(calcString)
if subtle.ConstantTimeCompare([]byte(claims["response"].(string)), []byte(calculated)) != 1 {
lg.Debug("invalid response", "path", r.URL.Path)
failedValidations.Inc()
clearCookie(w)
s.renderIndex(w, r)
return
}
slog.Debug("all checks passed")
r.Header.Add("X-Anubis-Status", "PASS-FULL")
s.rp.ServeHTTP(w, r)
}
func (s *Server) renderIndex(w http.ResponseWriter, r *http.Request) {
templ.Handler(
base("Making sure you're not a bot!", index()),
).ServeHTTP(w, r)
}
func (s *Server) makeChallenge(w http.ResponseWriter, r *http.Request) {
lg := slog.With("user_agent", r.UserAgent(), "accept_language", r.Header.Get("Accept-Language"), "priority", r.Header.Get("Priority"), "x-forwarded-for", r.Header.Get("X-Forwarded-For"), "x-real-ip", r.Header.Get("X-Real-Ip"))
cr, rule, err := s.check(r)
if err != nil {
lg.Error("check failed", "err", err)
w.WriteHeader(http.StatusInternalServerError)
json.NewEncoder(w).Encode(struct {
Error string `json:"error"`
}{
Error: "Internal Server Error: administrator has misconfigured Anubis. Please contact the administrator and ask them to look for the logs around \"makeChallenge\"",
})
return
}
lg = lg.With("check_result", cr)
challenge := s.challengeFor(r, rule.Challenge.Difficulty)
json.NewEncoder(w).Encode(struct {
Challenge string `json:"challenge"`
Rules *config.ChallengeRules `json:"rules"`
}{
Challenge: challenge,
Rules: rule.Challenge,
})
lg.Debug("made challenge", "challenge", challenge, "rules", rule.Challenge, "cr", cr)
challengesIssued.Inc()
}
func (s *Server) passChallenge(w http.ResponseWriter, r *http.Request) {
lg := slog.With(
"user_agent", r.UserAgent(),
"accept_language", r.Header.Get("Accept-Language"),
"priority", r.Header.Get("Priority"),
"x-forwarded-for", r.Header.Get("X-Forwarded-For"),
"x-real-ip", r.Header.Get("X-Real-Ip"),
)
cr, rule, err := s.check(r)
if err != nil {
lg.Error("check failed", "err", err)
templ.Handler(base("Oh noes!", errorPage("Internal Server Error: administrator has misconfigured Anubis. Please contact the administrator and ask them to look for the logs around \"passChallenge\".")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
lg = lg.With("check_result", cr)
nonceStr := r.FormValue("nonce")
if nonceStr == "" {
clearCookie(w)
lg.Debug("no nonce")
templ.Handler(base("Oh noes!", errorPage("missing nonce")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
elapsedTimeStr := r.FormValue("elapsedTime")
if elapsedTimeStr == "" {
clearCookie(w)
lg.Debug("no elapsedTime")
templ.Handler(base("Oh noes!", errorPage("missing elapsedTime")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
elapsedTime, err := strconv.ParseFloat(elapsedTimeStr, 64)
if err != nil {
clearCookie(w)
lg.Debug("elapsedTime doesn't parse", "err", err)
templ.Handler(base("Oh noes!", errorPage("invalid elapsedTime")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
lg.Info("challenge took", "elapsedTime", elapsedTime)
timeTaken.Observe(elapsedTime)
response := r.FormValue("response")
redir := r.FormValue("redir")
challenge := s.challengeFor(r, rule.Challenge.Difficulty)
nonce, err := strconv.Atoi(nonceStr)
if err != nil {
clearCookie(w)
lg.Debug("nonce doesn't parse", "err", err)
templ.Handler(base("Oh noes!", errorPage("invalid nonce")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
calcString := fmt.Sprintf("%s%d", challenge, nonce)
calculated := sha256sum(calcString)
if subtle.ConstantTimeCompare([]byte(response), []byte(calculated)) != 1 {
clearCookie(w)
lg.Debug("hash does not match", "got", response, "want", calculated)
templ.Handler(base("Oh noes!", errorPage("invalid response")), templ.WithStatus(http.StatusForbidden)).ServeHTTP(w, r)
failedValidations.Inc()
return
}
// compare the leading zeroes
if !strings.HasPrefix(response, strings.Repeat("0", *challengeDifficulty)) {
clearCookie(w)
lg.Debug("difficulty check failed", "response", response, "difficulty", *challengeDifficulty)
templ.Handler(base("Oh noes!", errorPage("invalid response")), templ.WithStatus(http.StatusForbidden)).ServeHTTP(w, r)
failedValidations.Inc()
return
}
// generate JWT cookie
token := jwt.NewWithClaims(jwt.SigningMethodEdDSA, jwt.MapClaims{
"challenge": challenge,
"nonce": nonce,
"response": response,
"iat": time.Now().Unix(),
"nbf": time.Now().Add(-1 * time.Minute).Unix(),
"exp": time.Now().Add(24 * 7 * time.Hour).Unix(),
})
tokenString, err := token.SignedString(s.priv)
if err != nil {
lg.Error("failed to sign JWT", "err", err)
clearCookie(w)
templ.Handler(base("Oh noes!", errorPage("failed to sign JWT")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
http.SetCookie(w, &http.Cookie{
Name: cookieName,
Value: tokenString,
Expires: time.Now().Add(24 * 7 * time.Hour),
SameSite: http.SameSiteLaxMode,
Path: "/",
})
challengesValidated.Inc()
lg.Debug("challenge passed, redirecting to app")
http.Redirect(w, r, redir, http.StatusFound)
}
func (s *Server) testError(w http.ResponseWriter, r *http.Request) {
err := r.FormValue("err")
templ.Handler(base("Oh noes!", errorPage(err)), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
}
func ohNoes(w http.ResponseWriter, r *http.Request, err error) {
slog.Error("super fatal error", "err", err)
templ.Handler(base("Oh noes!", errorPage("An internal server error happened")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
}
func clearCookie(w http.ResponseWriter) {
http.SetCookie(w, &http.Cookie{
Name: cookieName,
Value: "",
Expires: time.Now().Add(-1 * time.Hour),
MaxAge: -1,
SameSite: http.SameSiteLaxMode,
})
}
func randomJitter() bool {
return mrand.Intn(100) > 10
}
func serveMainJSWithBestEncoding(w http.ResponseWriter, r *http.Request) {
priorityList := []string{"zstd", "br", "gzip"}
enc2ext := map[string]string{
@@ -701,11 +279,11 @@ func serveMainJSWithBestEncoding(w http.ResponseWriter, r *http.Request) {
if strings.Contains(r.Header.Get("Accept-Encoding"), enc) {
w.Header().Set("Content-Type", "text/javascript")
w.Header().Set("Content-Encoding", enc)
http.ServeFileFS(w, r, static, "static/js/main.mjs."+enc2ext[enc])
http.ServeFileFS(w, r, web.Static, "static/js/main.mjs."+enc2ext[enc])
return
}
}
w.Header().Set("Content-Type", "text/javascript")
http.ServeFileFS(w, r, static, "static/js/main.mjs")
http.ServeFileFS(w, r, web.Static, "static/js/main.mjs")
}

View File

@@ -1,212 +0,0 @@
package main
import (
"encoding/json"
"errors"
"fmt"
"io"
"log"
"log/slog"
"net"
"net/http"
"regexp"
"github.com/TecharoHQ/anubis/cmd/anubis/internal/config"
"github.com/prometheus/client_golang/prometheus"
"github.com/prometheus/client_golang/prometheus/promauto"
"github.com/yl2chen/cidranger"
)
var (
policyApplications = promauto.NewCounterVec(prometheus.CounterOpts{
Name: "anubis_policy_results",
Help: "The results of each policy rule",
}, []string{"rule", "action"})
)
type ParsedConfig struct {
orig config.Config
Bots []Bot
DNSBL bool
}
type Bot struct {
Name string
UserAgent *regexp.Regexp
Path *regexp.Regexp
Action config.Rule `json:"action"`
Challenge *config.ChallengeRules
Ranger cidranger.Ranger
}
func (b Bot) Hash() (string, error) {
var pathRex string
if b.Path != nil {
pathRex = b.Path.String()
}
var userAgentRex string
if b.UserAgent != nil {
userAgentRex = b.UserAgent.String()
}
return sha256sum(fmt.Sprintf("%s::%s::%s", b.Name, pathRex, userAgentRex)), nil
}
func parseConfig(fin io.Reader, fname string, defaultDifficulty int) (*ParsedConfig, error) {
var c config.Config
if err := json.NewDecoder(fin).Decode(&c); err != nil {
return nil, fmt.Errorf("can't parse policy config JSON %s: %w", fname, err)
}
if err := c.Valid(); err != nil {
return nil, err
}
var err error
result := &ParsedConfig{
orig: c,
}
for _, b := range c.Bots {
if berr := b.Valid(); berr != nil {
err = errors.Join(err, berr)
continue
}
var botParseErr error
parsedBot := Bot{
Name: b.Name,
Action: b.Action,
}
if b.RemoteAddr != nil && len(b.RemoteAddr) > 0 {
parsedBot.Ranger = cidranger.NewPCTrieRanger()
for _, cidr := range b.RemoteAddr {
_, rng, err := net.ParseCIDR(cidr)
if err != nil {
return nil, fmt.Errorf("[unexpected] range %s not parsing: %w", cidr, err)
}
parsedBot.Ranger.Insert(cidranger.NewBasicRangerEntry(*rng))
}
}
if b.UserAgentRegex != nil {
userAgent, err := regexp.Compile(*b.UserAgentRegex)
if err != nil {
botParseErr = errors.Join(botParseErr, fmt.Errorf("while compiling user agent regexp: %w", err))
continue
} else {
parsedBot.UserAgent = userAgent
}
}
if b.PathRegex != nil {
path, err := regexp.Compile(*b.PathRegex)
if err != nil {
botParseErr = errors.Join(botParseErr, fmt.Errorf("while compiling path regexp: %w", err))
continue
} else {
parsedBot.Path = path
}
}
if b.Challenge == nil {
parsedBot.Challenge = &config.ChallengeRules{
Difficulty: defaultDifficulty,
ReportAs: defaultDifficulty,
Algorithm: config.AlgorithmFast,
}
} else {
parsedBot.Challenge = b.Challenge
if parsedBot.Challenge.Algorithm == config.AlgorithmUnknown {
parsedBot.Challenge.Algorithm = config.AlgorithmFast
}
}
result.Bots = append(result.Bots, parsedBot)
}
if err != nil {
return nil, fmt.Errorf("errors validating policy config JSON %s: %w", fname, err)
}
result.DNSBL = c.DNSBL
return result, nil
}
type CheckResult struct {
Name string
Rule config.Rule
}
func (cr CheckResult) LogValue() slog.Value {
return slog.GroupValue(
slog.String("name", cr.Name),
slog.String("rule", string(cr.Rule)))
}
func cr(name string, rule config.Rule) CheckResult {
return CheckResult{
Name: name,
Rule: rule,
}
}
func (s *Server) checkRemoteAddress(b Bot, addr net.IP) bool {
if b.Ranger == nil {
return false
}
ok, err := b.Ranger.Contains(addr)
if err != nil {
log.Panicf("[unexpected] something very funky is going on, %q does not have a calculable network number: %v", addr.String(), err)
}
return ok
}
// Check evaluates the list of rules, and returns the result
func (s *Server) check(r *http.Request) (CheckResult, *Bot, error) {
host := r.Header.Get("X-Real-Ip")
if host == "" {
return zilch[CheckResult](), nil, fmt.Errorf("[misconfiguration] X-Real-Ip header is not set")
}
addr := net.ParseIP(host)
if addr == nil {
return zilch[CheckResult](), nil, fmt.Errorf("[misconfiguration] %q is not an IP address", host)
}
for _, b := range s.policy.Bots {
if b.UserAgent != nil {
if uaMatch := b.UserAgent.MatchString(r.UserAgent()); uaMatch || (uaMatch && s.checkRemoteAddress(b, addr)) {
return cr("bot/"+b.Name, b.Action), &b, nil
}
}
if b.Path != nil {
if pathMatch := b.Path.MatchString(r.URL.Path); pathMatch || (pathMatch && s.checkRemoteAddress(b, addr)) {
return cr("bot/"+b.Name, b.Action), &b, nil
}
}
if b.Ranger != nil {
if s.checkRemoteAddress(b, addr) {
return cr("bot/"+b.Name, b.Action), &b, nil
}
}
}
return cr("default/allow", config.RuleAllow), &Bot{
Challenge: &config.ChallengeRules{
Difficulty: defaultDifficulty,
ReportAs: defaultDifficulty,
Algorithm: config.AlgorithmFast,
},
}, nil
}

View File

@@ -23,21 +23,8 @@ var (
githubEventName = flag.String("github-event-name", "", "GitHub event name")
pullRequestID = flag.Int("pull-request-id", -1, "GitHub pull request ID")
slogLevel = flag.String("slog-level", "INFO", "logging level (see https://pkg.go.dev/log/slog#hdr-Levels)")
knownContributors = []string{
"Xe",
}
)
func inList(needle string, haystack []string) bool {
for _, h := range haystack {
if h == needle {
return true
}
}
return false
}
func main() {
flagenv.Parse()
flag.Parse()
@@ -46,11 +33,7 @@ func main() {
koDockerRepo := strings.TrimRight(*dockerRepo, "/"+filepath.Base(*dockerRepo))
if *githubEventName == "pull_request" && !inList(*githubActor, knownContributors) {
if *pullRequestID == -1 {
log.Fatal("Must set --pull-request-id when --github-event-name=pull_request")
}
if *githubEventName == "pull_request" && *pullRequestID != -1 {
*dockerRepo = fmt.Sprintf("ttl.sh/techaro/pr-%d/anubis", *pullRequestID)
*dockerTags = fmt.Sprintf("ttl.sh/techaro/pr-%d/anubis:24h", *pullRequestID)
koDockerRepo = fmt.Sprintf("ttl.sh/techaro/pr-%d", *pullRequestID)

View File

@@ -7,7 +7,7 @@
},
{
"name": "googlebot",
"user_agent_regex": "\\+http\\:\\/\\/www\\.google\\.com/bot\\.html",
"user_agent_regex": "\\+http\\://www\\.google\\.com/bot\\.html",
"action": "ALLOW",
"remote_addresses": [
"2001:4860:4801:10::/64",
@@ -271,7 +271,7 @@
},
{
"name": "bingbot",
"user_agent_regex": "\\+http\\:\\/\\/www\\.bing\\.com/bingbot\\.htm",
"user_agent_regex": "\\+http\\://www\\.bing\\.com/bingbot\\.htm",
"action": "ALLOW",
"remote_addresses": [
"157.55.39.0/24",
@@ -306,7 +306,7 @@
},
{
"name": "qwantbot",
"user_agent_regex": "\\+https\\:\\/\\/help\\.qwant\\.com/bot/",
"user_agent_regex": "\\+https\\://help\\.qwant\\.com/bot/",
"action": "ALLOW",
"remote_addresses": [
"91.242.162.0/24"
@@ -314,7 +314,7 @@
},
{
"name": "kagibot",
"user_agent_regex": "\\+https\\:\\/\\/kagi\\.com/bot",
"user_agent_regex": "\\+https\\://kagi\\.com/bot",
"action": "ALLOW",
"remote_addresses": [
"216.18.205.234/32",
@@ -345,7 +345,7 @@
},
{
"name": "us-artificial-intelligence-scraper",
"user_agent_regex": "\\+https\\:\\/\\/github\\.com\\/US-Artificial-Intelligence\\/scraper",
"user_agent_regex": "\\+https\\://github\\.com/US-Artificial-Intelligence/scraper",
"action": "DENY"
},
{

8
data/embed.go Normal file
View File

@@ -0,0 +1,8 @@
package data
import "embed"
var (
//go:embed botPolicies.json
BotPolicies embed.FS
)

View File

@@ -1,17 +1,17 @@
package main
package decaymap
import (
"sync"
"time"
)
func zilch[T any]() T {
func Zilch[T any]() T {
var zero T
return zero
}
// DecayMap is a lazy key->value map. It's a wrapper around a map and a mutex. If values exceed their time-to-live, they are pruned at Get time.
type DecayMap[K comparable, V any] struct {
// Impl is a lazy key->value map. It's a wrapper around a map and a mutex. If values exceed their time-to-live, they are pruned at Get time.
type Impl[K comparable, V any] struct {
data map[K]decayMapEntry[V]
lock sync.RWMutex
}
@@ -21,17 +21,17 @@ type decayMapEntry[V any] struct {
expiry time.Time
}
// NewDecayMap creates a new DecayMap of key type K and value type V.
// New creates a new DecayMap of key type K and value type V.
//
// Key types must be comparable to work with maps.
func NewDecayMap[K comparable, V any]() *DecayMap[K, V] {
return &DecayMap[K, V]{
func New[K comparable, V any]() *Impl[K, V] {
return &Impl[K, V]{
data: make(map[K]decayMapEntry[V]),
}
}
// expire forcibly expires a key by setting its time-to-live one second in the past.
func (m *DecayMap[K, V]) expire(key K) bool {
func (m *Impl[K, V]) expire(key K) bool {
m.lock.RLock()
val, ok := m.data[key]
m.lock.RUnlock()
@@ -51,32 +51,32 @@ func (m *DecayMap[K, V]) expire(key K) bool {
// Get gets a value from the DecayMap by key.
//
// If a value has expired, forcibly delete it if it was not updated.
func (m *DecayMap[K, V]) Get(key K) (V, bool) {
func (m *Impl[K, V]) Get(key K) (V, bool) {
m.lock.RLock()
value, ok := m.data[key]
m.lock.RUnlock()
if !ok {
return zilch[V](), false
return Zilch[V](), false
}
if time.Now().After(value.expiry) {
m.lock.Lock()
// Since previously reading m.data[key], the value may have been updated.
// Delete the entry only if the expiry time is still the same.
if m.data[key].expiry == value.expiry {
if m.data[key].expiry.Equal(value.expiry) {
delete(m.data, key)
}
m.lock.Unlock()
return zilch[V](), false
return Zilch[V](), false
}
return value.Value, true
}
// Set sets a key value pair in the map.
func (m *DecayMap[K, V]) Set(key K, value V, ttl time.Duration) {
func (m *Impl[K, V]) Set(key K, value V, ttl time.Duration) {
m.lock.Lock()
defer m.lock.Unlock()

View File

@@ -1,12 +1,12 @@
package main
package decaymap
import (
"testing"
"time"
)
func TestDecayMap(t *testing.T) {
dm := NewDecayMap[string, string]()
func TestImpl(t *testing.T) {
dm := New[string, string]()
dm.Set("test", "hi", 5*time.Minute)

8
doc.go
View File

@@ -1,8 +0,0 @@
// Package Anubis contains the version number of Anubis.
package anubis
// Version is the current version of Anubis.
//
// This variable is set at build time using the -X linker flag. If not set,
// it defaults to "devel".
var Version = "devel"

View File

@@ -11,7 +11,15 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
## [Unreleased]
- ed25519 signing keys for Anubis can be stored in the flag `--ed25519-private-key-hex` or envvar `ED25519_PRIVATE_KEY_HEX`; if one is not provided when Anubis starts, a new one is generated and logged
- Fixed and clarified installation instructions
- Introduced integration tests using Playwright
- Refactor & Split up Anubis into cmd and lib.go
- Fixed bot check to only apply if address range matches
- Fix default difficulty setting that was broken in a refactor
- Linting fixes
- Extra cookie options can be set such as the name, domain, and partitioned flag [#73](https://github.com/TecharoHQ/anubis/issues/73)
- Make dark mode diff lines readable in the documentation
## v1.14.2
@@ -20,6 +28,7 @@ Livia sas Junius: Echo 2
- Remove default RSS reader rule as it may allow for a targeted attack against rails apps
[#67](https://github.com/TecharoHQ/anubis/pull/67)
- Whitelist MojeekBot in botPolicies [#47](https://github.com/TecharoHQ/anubis/issues/47)
- botPolicies regex has been cleaned up [#66](https://github.com/TecharoHQ/anubis/pull/66)
## v1.14.1

View File

@@ -2,8 +2,28 @@
title: Setting up Anubis
---
import RandomKey from "@site/src/components/RandomKey";
Anubis is meant to sit between your reverse proxy (such as Nginx or Caddy) and your target service. One instance of Anubis must be used per service you are protecting.
<center>
```mermaid
---
title: With Anubis installed
---
flowchart LR
LB(Load balancer /
TLS terminator)
Anubis(Anubis)
App(App)
LB --> Anubis --> App
```
</center>
Anubis is shipped in the Docker repo [`ghcr.io/techarohq/anubis`](https://github.com/TecharoHQ/anubis/pkgs/container/anubis). The following tags exist for your convenience:
| Tag | Meaning |
@@ -21,17 +41,33 @@ Anubis has very minimal system requirements. I suspect that 128Mi of ram may be
Anubis uses these environment variables for configuration:
| Environment Variable | Default value | Explanation |
| :--------------------- | :------------------------- | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `BIND` | `:8923` | The network address that Anubis listens on. For `unix`, set this to a path: `/run/anubis/instance.sock` |
| `BIND_NETWORK` | `tcp` | The address family that Anubis listens on. Accepts `tcp`, `unix` and anything Go's [`net.Listen`](https://pkg.go.dev/net#Listen) supports. |
| `DIFFICULTY` | `5` | The difficulty of the challenge, or the number of leading zeroes that must be in successful responses. |
| `METRICS_BIND` | `:9090` | The network address that Anubis serves Prometheus metrics on. See `BIND` for more information. |
| `METRICS_BIND_NETWORK` | `tcp` | The address family that the Anubis metrics server listens on. See `BIND_NETWORK` for more information. |
| `SOCKET_MODE` | `0770` | *Only used when at least one of the `*_BIND_NETWORK` variables are set to `unix`.* The socket mode (permissions) for Unix domain sockets. |
| `POLICY_FNAME` | unset | The file containing [bot policy configuration](./policies.md). See the bot policy documentation for more details. If unset, the default bot policy configuration is used. |
| `SERVE_ROBOTS_TXT` | `false` | If set `true`, Anubis will serve a default `robots.txt` file that disallows all known AI scrapers by name and then additionally disallows every scraper. This is useful if facts and circumstances make it difficult to change the underlying service to serve such a `robots.txt` file. |
| `TARGET` | `http://localhost:3923` | The URL of the service that Anubis should forward valid requests to. Supports Unix domain sockets, set this to a URI like so: `unix:///path/to/socket.sock`. |
| Environment Variable | Default value | Explanation |
| :------------------------ | :--------------------------------- | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `BIND` | `:8923` | The network address that Anubis listens on. For `unix`, set this to a path: `/run/anubis/instance.sock` |
| `BIND_NETWORK` | `tcp` | The address family that Anubis listens on. Accepts `tcp`, `unix` and anything Go's [`net.Listen`](https://pkg.go.dev/net#Listen) supports. |
| `COOKIE_DOMAIN` | unset | The domain the Anubis challenge pass cookie should be set to. This should be set to the domain you bought from your registrar (EG: `techaro.lol` if your webapp is running on `anubis.techaro.lol`). See [here](https://stackoverflow.com/a/1063760) for more information. |
| `COOKIE_NAME` | `within.website-x-cmd-anubis-auth` | The cookie that Anubis uses to determine if users have passed a challenge. This should not be changed without a good reason. |
| `COOKIE_PARTITIONED` | `false` | If set to `true`, enables the [partitioned (CHIPS) flag](https://developers.google.com/privacy-sandbox/cookies/chips), meaning that Anubis inside an iframe has a different set of cookies than the domain hosting the iframe. |
| `DIFFICULTY` | `5` | The difficulty of the challenge, or the number of leading zeroes that must be in successful responses. |
| `ED25519_PRIVATE_KEY_HEX` | unset | The hex-encoded ed25519 private key used to sign Anubis responses. If this is not set, Anubis will generate one for you. This should be exactly 64 characters long. See below for details. |
| `METRICS_BIND` | `:9090` | The network address that Anubis serves Prometheus metrics on. See `BIND` for more information. |
| `METRICS_BIND_NETWORK` | `tcp` | The address family that the Anubis metrics server listens on. See `BIND_NETWORK` for more information. |
| `SOCKET_MODE` | `0770` | _Only used when at least one of the `*_BIND_NETWORK` variables are set to `unix`._ The socket mode (permissions) for Unix domain sockets. |
| `POLICY_FNAME` | unset | The file containing [bot policy configuration](./policies.md). See the bot policy documentation for more details. If unset, the default bot policy configuration is used. |
| `SERVE_ROBOTS_TXT` | `false` | If set `true`, Anubis will serve a default `robots.txt` file that disallows all known AI scrapers by name and then additionally disallows every scraper. This is useful if facts and circumstances make it difficult to change the underlying service to serve such a `robots.txt` file. |
| `TARGET` | `http://localhost:3923` | The URL of the service that Anubis should forward valid requests to. Supports Unix domain sockets, set this to a URI like so: `unix:///path/to/socket.sock`. |
### Key generation
To generate an ed25519 private key, you can use this command:
```text
openssl rand -hex 32
```
Alternatively here is a key generated by your browser:
<RandomKey />
## Docker compose

View File

@@ -23,6 +23,6 @@ Anubis is a bit of a nuclear response. This will result in your website being bl
## Support
If you run into any issues running Anubis, please [open an issue](https://github.com/TecharoHQ/anubis/issues/new?template=Blank+issue) and tag it with the Anubis tag. Please include all the information I would need to diagnose your issue.
If you run into any issues running Anubis, please [open an issue](https://github.com/TecharoHQ/anubis/issues/new?template=Blank+issue) and include all the information I would need to diagnose your issue.
For live chat, please join the [Patreon](https://patreon.com/cadey) and ask in the Patron discord in the channel `#anubis`.

View File

@@ -22,7 +22,7 @@ spec:
ports:
- containerPort: 80
- name: anubis
image: ghcr.io/techarohq/anubis:latest
image: ghcr.io/techarohq/anubis:main
imagePullPolicy: Always
env:
- name: "BIND"

View File

@@ -0,0 +1,42 @@
import { useState, useCallback } from "react";
import Code from "@theme/CodeInline";
import BrowserOnly from "@docusaurus/BrowserOnly";
// https://www.xaymar.com/articles/2020/12/08/fastest-uint8array-to-hex-string-conversion-in-javascript/
function toHex(buffer) {
return Array.prototype.map
.call(buffer, (x) => ("00" + x.toString(16)).slice(-2))
.join("");
}
export const genRandomKey = (): String => {
const array = new Uint8Array(32);
self.crypto.getRandomValues(array);
return toHex(array);
};
export default function RandomKey() {
return (
<BrowserOnly fallback={<div>Loading...</div>}>
{() => {
const [key, setKey] = useState<String>(genRandomKey());
const genRandomKeyCb = useCallback(() => {
setKey(genRandomKey());
});
return (
<span>
<Code>{key}</Code>
<span style={{ marginLeft: "0.25rem", marginRight: "0.25rem" }} />
<button
onClick={() => {
genRandomKeyCb();
}}
>
</button>
</span>
);
}}
</BrowserOnly>
);
}

View File

@@ -15,6 +15,8 @@
--ifm-color-primary-lightest: #3cad6e;
--ifm-code-font-size: 95%;
--docusaurus-highlighted-code-line-bg: rgba(0, 0, 0, 0.1);
--code-block-diff-add-line-color: #ccffd8;
--code-block-diff-remove-line-color: #ffebe9;
}
/* For readability concerns, you should choose a lighter palette in dark mode. */
@@ -27,10 +29,12 @@
--ifm-color-primary-lighter: #32d8b4;
--ifm-color-primary-lightest: #4fddbf;
--docusaurus-highlighted-code-line-bg: rgba(0, 0, 0, 0.3);
--code-block-diff-add-line-color: #216932;
--code-block-diff-remove-line-color: #8b423b;
}
.code-block-diff-add-line {
background-color: #ccffd8;
background-color: var(--code-block-diff-add-line-color);
display: block;
margin: 0 -40px;
padding: 0 40px;
@@ -44,7 +48,7 @@
}
.code-block-diff-remove-line {
background-color: #ffebe9;
background-color: var(--code-block-diff-remove-line-color);
display: block;
margin: 0 -40px;
padding: 0 40px;

4
go.mod
View File

@@ -6,6 +6,7 @@ require (
github.com/a-h/templ v0.3.833
github.com/facebookgo/flagenv v0.0.0-20160425205200-fcd59fca7456
github.com/golang-jwt/jwt/v5 v5.2.2
github.com/playwright-community/playwright-go v0.5001.0
github.com/prometheus/client_golang v1.21.1
github.com/sebest/xff v0.0.0-20210106013422-671bd2870b3a
github.com/yl2chen/cidranger v1.0.2
@@ -21,11 +22,14 @@ require (
github.com/cespare/xxhash/v2 v2.3.0 // indirect
github.com/cli/browser v1.3.0 // indirect
github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc // indirect
github.com/deckarep/golang-set/v2 v2.6.0 // indirect
github.com/facebookgo/ensure v0.0.0-20160127193407-b4ab57deab51 // indirect
github.com/facebookgo/stack v0.0.0-20160209184415-751773369052 // indirect
github.com/facebookgo/subset v0.0.0-20150612182917-8dac2c3c4870 // indirect
github.com/fatih/color v1.16.0 // indirect
github.com/fsnotify/fsnotify v1.7.0 // indirect
github.com/go-jose/go-jose/v3 v3.0.4 // indirect
github.com/go-stack/stack v1.8.1 // indirect
github.com/klauspost/compress v1.17.11 // indirect
github.com/mattn/go-colorable v0.1.13 // indirect
github.com/mattn/go-isatty v0.0.20 // indirect

13
go.sum
View File

@@ -19,6 +19,8 @@ github.com/cli/browser v1.3.0/go.mod h1:HH8s+fOAxjhQoBUAsKuPCbqUuxZDhQ2/aD+SzsEf
github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc h1:U9qPSI2PIWSS1VwoXQT9A3Wy9MM3WgvqSxFWenqJduM=
github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
github.com/deckarep/golang-set/v2 v2.6.0 h1:XfcQbWM1LlMB8BsJ8N9vW5ehnnPVIw0je80NsVHagjM=
github.com/deckarep/golang-set/v2 v2.6.0/go.mod h1:VAky9rY/yGXJOLEDv3OMci+7wtDpOF4IN+y82NBOac4=
github.com/facebookgo/ensure v0.0.0-20160127193407-b4ab57deab51 h1:0JZ+dUmQeA8IIVUMzysrX4/AKuQwWhV2dYQuPZdvdSQ=
github.com/facebookgo/ensure v0.0.0-20160127193407-b4ab57deab51/go.mod h1:Yg+htXGokKKdzcwhuNDwVvN+uBxDGXJ7G/VN1d8fa64=
github.com/facebookgo/flagenv v0.0.0-20160425205200-fcd59fca7456 h1:CkmB2l68uhvRlwOTPrwnuitSxi/S3Cg4L5QYOcL9MBc=
@@ -31,8 +33,13 @@ github.com/fatih/color v1.16.0 h1:zmkK9Ngbjj+K0yRhTVONQh1p/HknKYSlNT+vZCzyokM=
github.com/fatih/color v1.16.0/go.mod h1:fL2Sau1YI5c0pdGEVCbKQbLXB6edEj1ZgiY4NijnWvE=
github.com/fsnotify/fsnotify v1.7.0 h1:8JEhPFa5W2WU7YfeZzPNqzMP6Lwt7L2715Ggo0nosvA=
github.com/fsnotify/fsnotify v1.7.0/go.mod h1:40Bi/Hjc2AVfZrqy+aj+yEI+/bRxZnMJyTJwOpGvigM=
github.com/go-jose/go-jose/v3 v3.0.4 h1:Wp5HA7bLQcKnf6YYao/4kpRpVMp/yf6+pJKV8WFSaNY=
github.com/go-jose/go-jose/v3 v3.0.4/go.mod h1:5b+7YgP7ZICgJDBdfjZaIt+H/9L9T/YQrVfLAMboGkQ=
github.com/go-stack/stack v1.8.1 h1:ntEHSVwIt7PNXNpgPmVfMrNhLtgjlmnZha2kOpuRiDw=
github.com/go-stack/stack v1.8.1/go.mod h1:dcoOX6HbPZSZptuspn9bctJ+N/CnF5gGygcUP3XYfe4=
github.com/golang-jwt/jwt/v5 v5.2.2 h1:Rl4B7itRWVtYIHFrSNd7vhTiz9UpLdi6gZhZ3wEeDy8=
github.com/golang-jwt/jwt/v5 v5.2.2/go.mod h1:pqrtFR0X4osieyHYxtmOUWsAWrfe1Q5UVIyoH402zdk=
github.com/google/go-cmp v0.5.9/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
github.com/google/go-cmp v0.6.0 h1:ofyhxvXcZhMsU5ulbFiLKl/XBFqE1GSq7atu8tAmTRI=
github.com/google/go-cmp v0.6.0/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
github.com/klauspost/compress v1.17.11 h1:In6xLpyWOi1+C7tXUUWv2ot1QvBjxevKAaI6IXrJmUc=
@@ -44,10 +51,14 @@ github.com/mattn/go-colorable v0.1.13/go.mod h1:7S9/ev0klgBDR4GtXTXX8a3vIGJpMovk
github.com/mattn/go-isatty v0.0.16/go.mod h1:kYGgaQfpe5nmfYZH+SKPsOc2e4SrIfOl2e/yFXSvRLM=
github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWEY=
github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
github.com/mitchellh/go-ps v1.0.0 h1:i6ampVEEF4wQFF+bkYfwYgY+F/uYJDktmvLPf7qIgjc=
github.com/mitchellh/go-ps v1.0.0/go.mod h1:J4lOc8z8yJs6vUwklHw2XEIiT4z4C40KtWVN3nvg8Pg=
github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 h1:C3w9PqII01/Oq1c1nUAm88MOHcQC9l5mIlSMApZMrHA=
github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822/go.mod h1:+n7T8mK8HuQTcFwEeznm/DIxMOiR9yIdICNftLE1DvQ=
github.com/natefinch/atomic v1.0.1 h1:ZPYKxkqQOx3KZ+RsbnP/YsgvxWQPGxjC0oBt2AhwV0A=
github.com/natefinch/atomic v1.0.1/go.mod h1:N/D/ELrljoqDyT3rZrsUmtsuzvHkeB/wWjHV22AZRbM=
github.com/playwright-community/playwright-go v0.5001.0 h1:EY3oB+rU9cUp6CLHguWE8VMZTwAg+83Yyb7dQqEmGLg=
github.com/playwright-community/playwright-go v0.5001.0/go.mod h1:kBNWs/w2aJ2ZUp1wEOOFLXgOqvppFngM5OS+qyhl+ZM=
github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2 h1:Jamvg5psRIccs7FGNTlIRMkT8wgtp5eCXdBlqhYGL6U=
github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
@@ -63,6 +74,7 @@ github.com/sebest/xff v0.0.0-20210106013422-671bd2870b3a h1:iLcLb5Fwwz7g/DLK89F+
github.com/sebest/xff v0.0.0-20210106013422-671bd2870b3a/go.mod h1:wozgYq9WEBQBaIJe4YZ0qTSFAMxmcwBhQH0fO0R34Z0=
github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
github.com/stretchr/testify v1.4.0/go.mod h1:j7eGeouHqKxXV5pUuKE4zz7dFj8WfuZ+81PSLYec5m4=
github.com/stretchr/testify v1.7.0/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
github.com/stretchr/testify v1.10.0 h1:Xv5erBjTwe/5IxqUQTdXv5kgmIvbHo3QQyRwhJsOfJA=
github.com/stretchr/testify v1.10.0/go.mod h1:r2ic/lqez/lEtzL7wO/rwa5dbSLXVDPFyf8C91i36aY=
github.com/yl2chen/cidranger v1.0.2 h1:lbOWZVCG1tCRX4u24kuM1Tb4nHqWkDxwLdoS+SevawU=
@@ -147,5 +159,6 @@ google.golang.org/protobuf v1.36.4 h1:6A3ZDJHn/eNqc1i+IdefRzy/9PokBTPvcqMySR7NNI
google.golang.org/protobuf v1.36.4/go.mod h1:9fA7Ob0pmnwhb644+1+CVWFRbNajQ6iRojtC/QF5bRE=
gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

12
internal/hash.go Normal file
View File

@@ -0,0 +1,12 @@
package internal
import (
"crypto/sha256"
"encoding/hex"
)
func SHA256sum(text string) string {
hash := sha256.New()
hash.Write([]byte(text))
return hex.EncodeToString(hash.Sum(nil))
}

View File

@@ -0,0 +1,435 @@
// Integration tests for Anubis, using Playwright.
//
// These tests require an already running Anubis and Playwright server.
//
// Anubis must be configured to redirect to the server started by the test suite.
// The bind address and the Anubis server can be specified using the flags `-bind` and `-anubis` respectively.
//
// Playwright must be started in server mode using `npx playwright@1.50.1 run-server --port 3000`.
// The version must match the minor used by the playwright-go package.
//
// On unsupported systems you may be able to use a container instead: https://playwright.dev/docs/docker#remote-connection
//
// In that case you may need to set the `-playwright` flag to the container's URL, and specify the `--host` the run-server command listens on.
package test
import (
"flag"
"fmt"
"net/http"
"net/http/httptest"
"net/url"
"os"
"os/exec"
"testing"
"time"
"github.com/TecharoHQ/anubis"
libanubis "github.com/TecharoHQ/anubis/lib"
"github.com/playwright-community/playwright-go"
)
var (
serverBindAddr = flag.String("bind", "localhost:3923", "test server bind address")
playwrightPort = flag.Int("playwright-port", 3000, "Playwright port")
playwrightServer = flag.String("playwright", "ws://localhost:3000", "Playwright server URL")
playwrightMaxTime = flag.Duration("playwright-max-time", 5*time.Second, "maximum time for Playwright requests")
playwrightMaxHardTime = flag.Duration("playwright-max-hard-time", 5*time.Minute, "maximum time for hard Playwright requests")
testCases = []testCase{
{
name: "firefox",
action: actionChallenge,
realIP: placeholderIP,
userAgent: "Mozilla/5.0 (X11; Linux x86_64; rv:136.0) Gecko/20100101 Firefox/136.0",
},
{
name: "headlessChrome",
action: actionDeny,
realIP: placeholderIP,
userAgent: "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) HeadlessChrome/120.0.6099.28 Safari/537.36",
},
{
name: "kagiBadIP",
action: actionChallenge,
isHard: true,
realIP: placeholderIP,
userAgent: "Mozilla/5.0 (compatible; Kagibot/1.0; +https://kagi.com/bot)",
},
{
name: "kagiGoodIP",
action: actionAllow,
realIP: "216.18.205.234",
userAgent: "Mozilla/5.0 (compatible; Kagibot/1.0; +https://kagi.com/bot)",
},
{
name: "unknownAgent",
action: actionAllow,
realIP: placeholderIP,
userAgent: "AnubisTest/0",
},
}
)
const (
actionAllow action = "ALLOW"
actionDeny action = "DENY"
actionChallenge action = "CHALLENGE"
placeholderIP = "fd11:5ee:bad:c0de::"
playwrightVersion = "1.50.1"
)
type action string
type testCase struct {
name string
action action
isHard bool
realIP, userAgent string
}
func doesNPXExist(t *testing.T) {
t.Helper()
if _, err := exec.LookPath("npx"); err != nil {
t.Skipf("npx not found in PATH, skipping integration smoke testing: %v", err)
}
}
func run(t *testing.T, command string) string {
t.Helper()
shPath, err := exec.LookPath("sh")
if err != nil {
t.Fatalf("[unexpected] %v", err)
}
t.Logf("running command: %s", command)
cmd := exec.Command(shPath, "-c", command)
cmd.Stdin = nil
cmd.Stderr = os.Stderr
output, err := cmd.Output()
if err != nil {
t.Fatalf("can't run command: %v", err)
}
return string(output)
}
func daemonize(t *testing.T, command string) {
t.Helper()
shPath, err := exec.LookPath("sh")
if err != nil {
t.Fatalf("[unexpected] %v", err)
}
t.Logf("daemonizing command: %s", command)
cmd := exec.Command(shPath, "-c", command)
cmd.Stdin = nil
cmd.Stderr = os.Stderr
cmd.Stdout = os.Stdout
if err := cmd.Start(); err != nil {
t.Fatalf("can't daemonize command: %v", err)
}
t.Cleanup(func() {
cmd.Process.Kill()
})
}
func startPlaywright(t *testing.T) {
t.Helper()
if os.Getenv("CI") == "true" {
run(t, fmt.Sprintf("npx --yes playwright@%s install --with-deps", playwrightVersion))
} else {
run(t, fmt.Sprintf("npx --yes playwright@%s install", playwrightVersion))
}
daemonize(t, fmt.Sprintf("npx --yes playwright@%s run-server --port %d", playwrightVersion, *playwrightPort))
for {
if _, err := http.Get(fmt.Sprintf("http://localhost:%d", *playwrightPort)); err != nil {
time.Sleep(500 * time.Millisecond)
continue
}
break
}
//nosleep:bypass XXX(Xe): Playwright doesn't have a good way to signal readiness. This is a HACK that will just let the tests pass.
time.Sleep(2 * time.Second)
}
func TestPlaywrightBrowser(t *testing.T) {
if os.Getenv("CI") == "true" {
t.Skip("XXX(Xe): This is broken in CI, will fix later")
}
if os.Getenv("DONT_USE_NETWORK") != "" {
t.Skip("test requires network egress")
return
}
doesNPXExist(t)
startPlaywright(t)
pw := setupPlaywright(t)
anubisURL := spawnAnubis(t)
browsers := []playwright.BrowserType{pw.Chromium, pw.Firefox, pw.WebKit}
for _, typ := range browsers {
t.Run(typ.Name()+"/warmup", func(t *testing.T) {
browser, err := typ.Connect(buildBrowserConnect(typ.Name()), playwright.BrowserTypeConnectOptions{
ExposeNetwork: playwright.String("<loopback>"),
})
if err != nil {
t.Fatalf("could not connect to remote browser: %v", err)
}
defer browser.Close()
ctx, err := browser.NewContext(playwright.BrowserNewContextOptions{
AcceptDownloads: playwright.Bool(false),
ExtraHttpHeaders: map[string]string{
"X-Real-Ip": "127.0.0.1",
},
UserAgent: playwright.String("Sephiroth"),
})
if err != nil {
t.Fatalf("could not create context: %v", err)
}
defer ctx.Close()
page, err := ctx.NewPage()
if err != nil {
t.Fatalf("could not create page: %v", err)
}
defer page.Close()
timeout := 2.0
page.Goto(anubisURL, playwright.PageGotoOptions{
Timeout: &timeout,
})
})
for _, tc := range testCases {
name := fmt.Sprintf("%s/%s", typ.Name(), tc.name)
t.Run(name, func(t *testing.T) {
_, hasDeadline := t.Deadline()
if tc.isHard && hasDeadline {
t.Skip("skipping hard challenge with deadline")
}
perfomedAction := executeTestCase(t, tc, typ, anubisURL)
if perfomedAction != tc.action {
t.Errorf("unexpected test result, expected %s, got %s", tc.action, perfomedAction)
} else {
t.Logf("test passed")
}
})
}
}
}
func buildBrowserConnect(name string) string {
u, _ := url.Parse(*playwrightServer)
q := u.Query()
q.Set("browser", name)
u.RawQuery = q.Encode()
return u.String()
}
func executeTestCase(t *testing.T, tc testCase, typ playwright.BrowserType, anubisURL string) action {
deadline, _ := t.Deadline()
browser, err := typ.Connect(buildBrowserConnect(typ.Name()), playwright.BrowserTypeConnectOptions{
ExposeNetwork: playwright.String("<loopback>"),
})
if err != nil {
t.Fatalf("could not connect to remote browser: %v", err)
}
defer browser.Close()
ctx, err := browser.NewContext(playwright.BrowserNewContextOptions{
AcceptDownloads: playwright.Bool(false),
ExtraHttpHeaders: map[string]string{
"X-Real-Ip": tc.realIP,
},
UserAgent: playwright.String(tc.userAgent),
})
if err != nil {
t.Fatalf("could not create context: %v", err)
}
defer ctx.Close()
page, err := ctx.NewPage()
if err != nil {
t.Fatalf("could not create page: %v", err)
}
defer page.Close()
// Attempt challenge.
start := time.Now()
_, err = page.Goto(anubisURL, playwright.PageGotoOptions{
Timeout: pwTimeout(tc, deadline),
})
if err != nil {
pwFail(t, page, "could not navigate to test server: %v", err)
}
hadChallenge := false
switch tc.action {
case actionChallenge:
// FIXME: This could race if challenge is completed too quickly.
checkImage(t, tc, deadline, page, "#image[src*=pensive], #image[src*=happy]")
hadChallenge = true
case actionDeny:
checkImage(t, tc, deadline, page, "#image[src*=sad]")
return actionDeny
}
// Ensure protected resource was provided.
res, err := page.Locator("#anubis-test").TextContent(playwright.LocatorTextContentOptions{
Timeout: pwTimeout(tc, deadline),
})
end := time.Now()
if err != nil {
pwFail(t, page, "could not get text content: %v", err)
}
var tm int64
if _, err := fmt.Sscanf(res, "%d", &tm); err != nil {
pwFail(t, page, "unexpected output: %s", res)
}
if tm < start.Unix() || end.Unix() < tm {
pwFail(t, page, "unexpected timestamp in output: %d not in range %d..%d", tm, start.Unix(), end.Unix())
}
if hadChallenge {
return actionChallenge
} else {
return actionAllow
}
}
func checkImage(t *testing.T, tc testCase, deadline time.Time, page playwright.Page, locator string) {
image := page.Locator(locator)
err := image.WaitFor(playwright.LocatorWaitForOptions{
Timeout: pwTimeout(tc, deadline),
})
if err != nil {
pwFail(t, page, "could not wait for result: %v", err)
}
failIsVisible, err := image.IsVisible()
if err != nil {
pwFail(t, page, "could not check result image: %v", err)
}
if !failIsVisible {
pwFail(t, page, "expected result image not visible")
}
}
func pwFail(t *testing.T, page playwright.Page, format string, args ...any) {
t.Helper()
saveScreenshot(t, page)
t.Fatalf(format, args...)
}
func pwTimeout(tc testCase, deadline time.Time) *float64 {
max := *playwrightMaxTime
if tc.isHard {
max = *playwrightMaxHardTime
}
d := time.Until(deadline)
if d <= 0 || d > max {
return playwright.Float(float64(max.Milliseconds()))
}
return playwright.Float(float64(d.Milliseconds()))
}
func saveScreenshot(t *testing.T, page playwright.Page) {
t.Helper()
data, err := page.Screenshot()
if err != nil {
t.Logf("could not take screenshot: %v", err)
return
}
f, err := os.CreateTemp("", "anubis-test-fail-*.png")
if err != nil {
t.Logf("could not create temporary file: %v", err)
return
}
defer f.Close()
_, err = f.Write(data)
if err != nil {
t.Logf("could not write screenshot: %v", err)
return
}
t.Logf("screenshot saved to %s", f.Name())
}
func setupPlaywright(t *testing.T) *playwright.Playwright {
err := playwright.Install(&playwright.RunOptions{
SkipInstallBrowsers: true,
})
if err != nil {
t.Fatalf("could not install Playwright: %v", err)
}
pw, err := playwright.Run()
if err != nil {
t.Fatalf("could not start Playwright: %v", err)
}
return pw
}
func spawnAnubis(t *testing.T) string {
t.Helper()
h := http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
w.Header().Add("Content-Type", "text/html")
fmt.Fprintf(w, "<html><body><span id=anubis-test>%d</span></body></html>", time.Now().Unix())
})
policy, err := libanubis.LoadPoliciesOrDefault("", anubis.DefaultDifficulty)
if err != nil {
t.Fatal(err)
}
s, err := libanubis.New(libanubis.Options{
Next: h,
Policy: policy,
ServeRobotsTXT: true,
})
if err != nil {
t.Fatalf("can't construct libanubis.Server: %v", err)
}
ts := httptest.NewServer(s)
t.Log(ts.URL)
t.Cleanup(func() {
ts.Close()
})
return ts.URL
}

527
lib/anubis.go Normal file
View File

@@ -0,0 +1,527 @@
package lib
import (
"crypto/ed25519"
"crypto/rand"
"crypto/sha256"
"crypto/subtle"
"encoding/json"
"fmt"
"io"
"log"
"log/slog"
"math"
"net"
"net/http"
"os"
"strconv"
"strings"
"time"
"github.com/a-h/templ"
"github.com/golang-jwt/jwt/v5"
"github.com/prometheus/client_golang/prometheus"
"github.com/prometheus/client_golang/prometheus/promauto"
"github.com/TecharoHQ/anubis"
"github.com/TecharoHQ/anubis/data"
"github.com/TecharoHQ/anubis/decaymap"
"github.com/TecharoHQ/anubis/internal"
"github.com/TecharoHQ/anubis/internal/dnsbl"
"github.com/TecharoHQ/anubis/lib/policy"
"github.com/TecharoHQ/anubis/lib/policy/config"
"github.com/TecharoHQ/anubis/web"
"github.com/TecharoHQ/anubis/xess"
)
var (
challengesIssued = promauto.NewCounter(prometheus.CounterOpts{
Name: "anubis_challenges_issued",
Help: "The total number of challenges issued",
})
challengesValidated = promauto.NewCounter(prometheus.CounterOpts{
Name: "anubis_challenges_validated",
Help: "The total number of challenges validated",
})
droneBLHits = promauto.NewCounterVec(prometheus.CounterOpts{
Name: "anubis_dronebl_hits",
Help: "The total number of hits from DroneBL",
}, []string{"status"})
failedValidations = promauto.NewCounter(prometheus.CounterOpts{
Name: "anubis_failed_validations",
Help: "The total number of failed validations",
})
timeTaken = promauto.NewHistogram(prometheus.HistogramOpts{
Name: "anubis_time_taken",
Help: "The time taken for a browser to generate a response (milliseconds)",
Buckets: prometheus.ExponentialBucketsRange(1, math.Pow(2, 18), 19),
})
)
type Options struct {
Next http.Handler
Policy *policy.ParsedConfig
ServeRobotsTXT bool
PrivateKey ed25519.PrivateKey
CookieDomain string
CookieName string
CookiePartitioned bool
}
func LoadPoliciesOrDefault(fname string, defaultDifficulty int) (*policy.ParsedConfig, error) {
var fin io.ReadCloser
var err error
if fname != "" {
fin, err = os.Open(fname)
if err != nil {
return nil, fmt.Errorf("can't parse policy file %s: %w", fname, err)
}
} else {
fname = "(data)/botPolicies.json"
fin, err = data.BotPolicies.Open("botPolicies.json")
if err != nil {
return nil, fmt.Errorf("[unexpected] can't parse builtin policy file %s: %w", fname, err)
}
}
defer fin.Close()
policy, err := policy.ParseConfig(fin, fname, defaultDifficulty)
return policy, err
}
func New(opts Options) (*Server, error) {
if opts.PrivateKey == nil {
slog.Debug("opts.PrivateKey not set, generating a new one")
_, priv, err := ed25519.GenerateKey(rand.Reader)
if err != nil {
return nil, fmt.Errorf("lib: can't generate private key: %v", err)
}
opts.PrivateKey = priv
}
result := &Server{
next: opts.Next,
priv: opts.PrivateKey,
pub: opts.PrivateKey.Public().(ed25519.PublicKey),
policy: opts.Policy,
opts: opts,
DNSBLCache: decaymap.New[string, dnsbl.DroneBLResponse](),
}
mux := http.NewServeMux()
xess.Mount(mux)
mux.Handle(anubis.StaticPath, internal.UnchangingCache(http.StripPrefix(anubis.StaticPath, http.FileServerFS(web.Static))))
if opts.ServeRobotsTXT {
mux.HandleFunc("/robots.txt", func(w http.ResponseWriter, r *http.Request) {
http.ServeFileFS(w, r, web.Static, "static/robots.txt")
})
mux.HandleFunc("/.well-known/robots.txt", func(w http.ResponseWriter, r *http.Request) {
http.ServeFileFS(w, r, web.Static, "static/robots.txt")
})
}
// mux.HandleFunc("GET /.within.website/x/cmd/anubis/static/js/main.mjs", serveMainJSWithBestEncoding)
mux.HandleFunc("POST /.within.website/x/cmd/anubis/api/make-challenge", result.MakeChallenge)
mux.HandleFunc("GET /.within.website/x/cmd/anubis/api/pass-challenge", result.PassChallenge)
mux.HandleFunc("GET /.within.website/x/cmd/anubis/api/test-error", result.TestError)
mux.HandleFunc("/", result.MaybeReverseProxy)
result.mux = mux
return result, nil
}
type Server struct {
mux *http.ServeMux
next http.Handler
priv ed25519.PrivateKey
pub ed25519.PublicKey
policy *policy.ParsedConfig
opts Options
DNSBLCache *decaymap.Impl[string, dnsbl.DroneBLResponse]
ChallengeDifficulty int
}
func (s *Server) ServeHTTP(w http.ResponseWriter, r *http.Request) {
s.mux.ServeHTTP(w, r)
}
func (s *Server) challengeFor(r *http.Request, difficulty int) string {
fp := sha256.Sum256(s.priv.Seed())
data := fmt.Sprintf(
"Accept-Language=%s,X-Real-IP=%s,User-Agent=%s,WeekTime=%s,Fingerprint=%x,Difficulty=%d",
r.Header.Get("Accept-Language"),
r.Header.Get("X-Real-Ip"),
r.UserAgent(),
time.Now().UTC().Round(24*7*time.Hour).Format(time.RFC3339),
fp,
difficulty,
)
return internal.SHA256sum(data)
}
func (s *Server) MaybeReverseProxy(w http.ResponseWriter, r *http.Request) {
lg := slog.With(
"user_agent", r.UserAgent(),
"accept_language", r.Header.Get("Accept-Language"),
"priority", r.Header.Get("Priority"),
"x-forwarded-for",
r.Header.Get("X-Forwarded-For"),
"x-real-ip", r.Header.Get("X-Real-Ip"),
)
cr, rule, err := s.check(r)
if err != nil {
lg.Error("check failed", "err", err)
templ.Handler(web.Base("Oh noes!", web.ErrorPage("Internal Server Error: administrator has misconfigured Anubis. Please contact the administrator and ask them to look for the logs around \"maybeReverseProxy\"")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
r.Header.Add("X-Anubis-Rule", cr.Name)
r.Header.Add("X-Anubis-Action", string(cr.Rule))
lg = lg.With("check_result", cr)
policy.PolicyApplications.WithLabelValues(cr.Name, string(cr.Rule)).Add(1)
ip := r.Header.Get("X-Real-Ip")
if s.policy.DNSBL && ip != "" {
resp, ok := s.DNSBLCache.Get(ip)
if !ok {
lg.Debug("looking up ip in dnsbl")
resp, err := dnsbl.Lookup(ip)
if err != nil {
lg.Error("can't look up ip in dnsbl", "err", err)
}
s.DNSBLCache.Set(ip, resp, 24*time.Hour)
droneBLHits.WithLabelValues(resp.String()).Inc()
}
if resp != dnsbl.AllGood {
lg.Info("DNSBL hit", "status", resp.String())
templ.Handler(web.Base("Oh noes!", web.ErrorPage(fmt.Sprintf("DroneBL reported an entry: %s, see https://dronebl.org/lookup?ip=%s", resp.String(), ip))), templ.WithStatus(http.StatusOK)).ServeHTTP(w, r)
return
}
}
switch cr.Rule {
case config.RuleAllow:
lg.Debug("allowing traffic to origin (explicit)")
s.next.ServeHTTP(w, r)
return
case config.RuleDeny:
s.ClearCookie(w)
lg.Info("explicit deny")
if rule == nil {
lg.Error("rule is nil, cannot calculate checksum")
templ.Handler(web.Base("Oh noes!", web.ErrorPage("Other internal server error (contact the admin)")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
hash, err := rule.Hash()
if err != nil {
lg.Error("can't calculate checksum of rule", "err", err)
templ.Handler(web.Base("Oh noes!", web.ErrorPage("Other internal server error (contact the admin)")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
lg.Debug("rule hash", "hash", hash)
templ.Handler(web.Base("Oh noes!", web.ErrorPage(fmt.Sprintf("Access Denied: error code %s", hash))), templ.WithStatus(http.StatusOK)).ServeHTTP(w, r)
return
case config.RuleChallenge:
lg.Debug("challenge requested")
default:
s.ClearCookie(w)
templ.Handler(web.Base("Oh noes!", web.ErrorPage("Other internal server error (contact the admin)")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
ckie, err := r.Cookie(s.opts.CookieName)
if err != nil {
lg.Debug("cookie not found", "path", r.URL.Path)
s.ClearCookie(w)
s.RenderIndex(w, r)
return
}
if err := ckie.Valid(); err != nil {
lg.Debug("cookie is invalid", "err", err)
s.ClearCookie(w)
s.RenderIndex(w, r)
return
}
if time.Now().After(ckie.Expires) && !ckie.Expires.IsZero() {
lg.Debug("cookie expired", "path", r.URL.Path)
s.ClearCookie(w)
s.RenderIndex(w, r)
return
}
token, err := jwt.ParseWithClaims(ckie.Value, jwt.MapClaims{}, func(token *jwt.Token) (interface{}, error) {
return s.pub, nil
}, jwt.WithExpirationRequired(), jwt.WithStrictDecoding())
if err != nil || !token.Valid {
lg.Debug("invalid token", "path", r.URL.Path, "err", err)
s.ClearCookie(w)
s.RenderIndex(w, r)
return
}
if randomJitter() {
r.Header.Add("X-Anubis-Status", "PASS-BRIEF")
lg.Debug("cookie is not enrolled into secondary screening")
s.next.ServeHTTP(w, r)
return
}
claims, ok := token.Claims.(jwt.MapClaims)
if !ok {
lg.Debug("invalid token claims type", "path", r.URL.Path)
s.ClearCookie(w)
s.RenderIndex(w, r)
return
}
challenge := s.challengeFor(r, rule.Challenge.Difficulty)
if claims["challenge"] != challenge {
lg.Debug("invalid challenge", "path", r.URL.Path)
s.ClearCookie(w)
s.RenderIndex(w, r)
return
}
var nonce int
if v, ok := claims["nonce"].(float64); ok {
nonce = int(v)
}
calcString := fmt.Sprintf("%s%d", challenge, nonce)
calculated := internal.SHA256sum(calcString)
if subtle.ConstantTimeCompare([]byte(claims["response"].(string)), []byte(calculated)) != 1 {
lg.Debug("invalid response", "path", r.URL.Path)
failedValidations.Inc()
s.ClearCookie(w)
s.RenderIndex(w, r)
return
}
slog.Debug("all checks passed")
r.Header.Add("X-Anubis-Status", "PASS-FULL")
s.next.ServeHTTP(w, r)
}
func (s *Server) RenderIndex(w http.ResponseWriter, r *http.Request) {
templ.Handler(
web.Base("Making sure you're not a bot!", web.Index()),
).ServeHTTP(w, r)
}
func (s *Server) MakeChallenge(w http.ResponseWriter, r *http.Request) {
lg := slog.With("user_agent", r.UserAgent(), "accept_language", r.Header.Get("Accept-Language"), "priority", r.Header.Get("Priority"), "x-forwarded-for", r.Header.Get("X-Forwarded-For"), "x-real-ip", r.Header.Get("X-Real-Ip"))
cr, rule, err := s.check(r)
if err != nil {
lg.Error("check failed", "err", err)
w.WriteHeader(http.StatusInternalServerError)
json.NewEncoder(w).Encode(struct {
Error string `json:"error"`
}{
Error: "Internal Server Error: administrator has misconfigured Anubis. Please contact the administrator and ask them to look for the logs around \"makeChallenge\"",
})
return
}
lg = lg.With("check_result", cr)
challenge := s.challengeFor(r, rule.Challenge.Difficulty)
json.NewEncoder(w).Encode(struct {
Challenge string `json:"challenge"`
Rules *config.ChallengeRules `json:"rules"`
}{
Challenge: challenge,
Rules: rule.Challenge,
})
lg.Debug("made challenge", "challenge", challenge, "rules", rule.Challenge, "cr", cr)
challengesIssued.Inc()
}
func (s *Server) PassChallenge(w http.ResponseWriter, r *http.Request) {
lg := slog.With(
"user_agent", r.UserAgent(),
"accept_language", r.Header.Get("Accept-Language"),
"priority", r.Header.Get("Priority"),
"x-forwarded-for", r.Header.Get("X-Forwarded-For"),
"x-real-ip", r.Header.Get("X-Real-Ip"),
)
cr, rule, err := s.check(r)
if err != nil {
lg.Error("check failed", "err", err)
templ.Handler(web.Base("Oh noes!", web.ErrorPage("Internal Server Error: administrator has misconfigured Anubis. Please contact the administrator and ask them to look for the logs around \"passChallenge\".")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
lg = lg.With("check_result", cr)
nonceStr := r.FormValue("nonce")
if nonceStr == "" {
s.ClearCookie(w)
lg.Debug("no nonce")
templ.Handler(web.Base("Oh noes!", web.ErrorPage("missing nonce")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
elapsedTimeStr := r.FormValue("elapsedTime")
if elapsedTimeStr == "" {
s.ClearCookie(w)
lg.Debug("no elapsedTime")
templ.Handler(web.Base("Oh noes!", web.ErrorPage("missing elapsedTime")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
elapsedTime, err := strconv.ParseFloat(elapsedTimeStr, 64)
if err != nil {
s.ClearCookie(w)
lg.Debug("elapsedTime doesn't parse", "err", err)
templ.Handler(web.Base("Oh noes!", web.ErrorPage("invalid elapsedTime")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
lg.Info("challenge took", "elapsedTime", elapsedTime)
timeTaken.Observe(elapsedTime)
response := r.FormValue("response")
redir := r.FormValue("redir")
challenge := s.challengeFor(r, rule.Challenge.Difficulty)
nonce, err := strconv.Atoi(nonceStr)
if err != nil {
s.ClearCookie(w)
lg.Debug("nonce doesn't parse", "err", err)
templ.Handler(web.Base("Oh noes!", web.ErrorPage("invalid nonce")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
calcString := fmt.Sprintf("%s%d", challenge, nonce)
calculated := internal.SHA256sum(calcString)
if subtle.ConstantTimeCompare([]byte(response), []byte(calculated)) != 1 {
s.ClearCookie(w)
lg.Debug("hash does not match", "got", response, "want", calculated)
templ.Handler(web.Base("Oh noes!", web.ErrorPage("invalid response")), templ.WithStatus(http.StatusForbidden)).ServeHTTP(w, r)
failedValidations.Inc()
return
}
// compare the leading zeroes
if !strings.HasPrefix(response, strings.Repeat("0", s.ChallengeDifficulty)) {
s.ClearCookie(w)
lg.Debug("difficulty check failed", "response", response, "difficulty", s.ChallengeDifficulty)
templ.Handler(web.Base("Oh noes!", web.ErrorPage("invalid response")), templ.WithStatus(http.StatusForbidden)).ServeHTTP(w, r)
failedValidations.Inc()
return
}
// generate JWT cookie
token := jwt.NewWithClaims(jwt.SigningMethodEdDSA, jwt.MapClaims{
"challenge": challenge,
"nonce": nonce,
"response": response,
"iat": time.Now().Unix(),
"nbf": time.Now().Add(-1 * time.Minute).Unix(),
"exp": time.Now().Add(24 * 7 * time.Hour).Unix(),
})
tokenString, err := token.SignedString(s.priv)
if err != nil {
lg.Error("failed to sign JWT", "err", err)
s.ClearCookie(w)
templ.Handler(web.Base("Oh noes!", web.ErrorPage("failed to sign JWT")), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
return
}
http.SetCookie(w, &http.Cookie{
Name: s.opts.CookieName,
Value: tokenString,
Expires: time.Now().Add(24 * 7 * time.Hour),
Domain: s.opts.CookieDomain,
Partitioned: s.opts.CookiePartitioned,
SameSite: http.SameSiteLaxMode,
})
challengesValidated.Inc()
lg.Debug("challenge passed, redirecting to app")
http.Redirect(w, r, redir, http.StatusFound)
}
func (s *Server) TestError(w http.ResponseWriter, r *http.Request) {
err := r.FormValue("err")
templ.Handler(web.Base("Oh noes!", web.ErrorPage(err)), templ.WithStatus(http.StatusInternalServerError)).ServeHTTP(w, r)
}
// Check evaluates the list of rules, and returns the result
func (s *Server) check(r *http.Request) (CheckResult, *policy.Bot, error) {
host := r.Header.Get("X-Real-Ip")
if host == "" {
return decaymap.Zilch[CheckResult](), nil, fmt.Errorf("[misconfiguration] X-Real-Ip header is not set")
}
addr := net.ParseIP(host)
if addr == nil {
return decaymap.Zilch[CheckResult](), nil, fmt.Errorf("[misconfiguration] %q is not an IP address", host)
}
for _, b := range s.policy.Bots {
if b.UserAgent != nil {
if b.UserAgent.MatchString(r.UserAgent()) && s.checkRemoteAddress(b, addr) {
return cr("bot/"+b.Name, b.Action), &b, nil
}
}
if b.Path != nil {
if b.Path.MatchString(r.URL.Path) && s.checkRemoteAddress(b, addr) {
return cr("bot/"+b.Name, b.Action), &b, nil
}
}
if b.Ranger != nil {
if s.checkRemoteAddress(b, addr) {
return cr("bot/"+b.Name, b.Action), &b, nil
}
}
}
return cr("default/allow", config.RuleAllow), &policy.Bot{
Challenge: &config.ChallengeRules{
Difficulty: s.policy.DefaultDifficulty,
ReportAs: s.policy.DefaultDifficulty,
Algorithm: config.AlgorithmFast,
},
}, nil
}
func (s *Server) checkRemoteAddress(b policy.Bot, addr net.IP) bool {
if b.Ranger == nil {
return true
}
ok, err := b.Ranger.Contains(addr)
if err != nil {
log.Panicf("[unexpected] something very funky is going on, %q does not have a calculable network number: %v", addr.String(), err)
}
return ok
}

158
lib/anubis_test.go Normal file
View File

@@ -0,0 +1,158 @@
package lib
import (
"encoding/json"
"fmt"
"net/http"
"net/http/httptest"
"testing"
"github.com/TecharoHQ/anubis"
"github.com/TecharoHQ/anubis/internal"
"github.com/TecharoHQ/anubis/lib/policy"
)
func loadPolicies(t *testing.T, fname string) *policy.ParsedConfig {
t.Helper()
policy, err := LoadPoliciesOrDefault("", anubis.DefaultDifficulty)
if err != nil {
t.Fatal(err)
}
return policy
}
func spawnAnubis(t *testing.T, opts Options) *Server {
t.Helper()
s, err := New(opts)
if err != nil {
t.Fatalf("can't construct libanubis.Server: %v", err)
}
return s
}
func TestCookieSettings(t *testing.T) {
pol := loadPolicies(t, "")
pol.DefaultDifficulty = 0
srv := spawnAnubis(t, Options{
Next: http.NewServeMux(),
Policy: pol,
CookieDomain: "local.cetacean.club",
CookiePartitioned: true,
CookieName: t.Name(),
})
ts := httptest.NewServer(internal.DefaultXRealIP("127.0.0.1", srv))
defer ts.Close()
cli := &http.Client{
CheckRedirect: func(req *http.Request, via []*http.Request) error {
return http.ErrUseLastResponse
},
}
resp, err := cli.Post(ts.URL+"/.within.website/x/cmd/anubis/api/make-challenge", "", nil)
if err != nil {
t.Fatalf("can't request challenge: %v", err)
}
defer resp.Body.Close()
var chall = struct {
Challenge string `json:"challenge"`
}{}
if err := json.NewDecoder(resp.Body).Decode(&chall); err != nil {
t.Fatalf("can't read challenge response body: %v", err)
}
nonce := 0
elapsedTime := 420
redir := "/"
calcString := fmt.Sprintf("%s%d", chall.Challenge, nonce)
calculated := internal.SHA256sum(calcString)
req, err := http.NewRequest(http.MethodGet, ts.URL+"/.within.website/x/cmd/anubis/api/pass-challenge", nil)
if err != nil {
t.Fatalf("can't make request: %v", err)
}
q := req.URL.Query()
q.Set("response", calculated)
q.Set("nonce", fmt.Sprint(nonce))
q.Set("redir", redir)
q.Set("elapsedTime", fmt.Sprint(elapsedTime))
req.URL.RawQuery = q.Encode()
resp, err = cli.Do(req)
if err != nil {
t.Fatalf("can't do challenge passing")
}
if resp.StatusCode != http.StatusFound {
t.Errorf("wanted %d, got: %d", http.StatusFound, resp.StatusCode)
}
found := false
for _, cookie := range resp.Cookies() {
t.Logf("%#v", cookie)
if cookie.Name == t.Name() {
found = true
}
if found && cookie.Domain != "local.cetacean.club" {
t.Errorf("cookie domain is wrong, wanted local.cetacean.club, got: %s", cookie.Domain)
}
}
if !found {
t.Errorf("Cookie %q not found", t.Name())
}
}
func TestCheckDefaultDifficultyMatchesPolicy(t *testing.T) {
h := http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
fmt.Fprintln(w, "OK")
})
for i := 1; i < 10; i++ {
t.Run(fmt.Sprint(i), func(t *testing.T) {
policy, err := LoadPoliciesOrDefault("", i)
if err != nil {
t.Fatal(err)
}
s, err := New(Options{
Next: h,
Policy: policy,
ServeRobotsTXT: true,
})
if err != nil {
t.Fatalf("can't construct libanubis.Server: %v", err)
}
req, err := http.NewRequest(http.MethodGet, "/", nil)
if err != nil {
t.Fatal(err)
}
req.Header.Add("X-Real-Ip", "127.0.0.1")
_, bot, err := s.check(req)
if err != nil {
t.Fatal(err)
}
if bot.Challenge.Difficulty != i {
t.Errorf("Challenge.Difficulty is wrong, wanted %d, got: %d", i, bot.Challenge.Difficulty)
}
if bot.Challenge.ReportAs != i {
t.Errorf("Challenge.ReportAs is wrong, wanted %d, got: %d", i, bot.Challenge.ReportAs)
}
})
}
}

25
lib/checkresult.go Normal file
View File

@@ -0,0 +1,25 @@
package lib
import (
"log/slog"
"github.com/TecharoHQ/anubis/lib/policy/config"
)
type CheckResult struct {
Name string
Rule config.Rule
}
func (cr CheckResult) LogValue() slog.Value {
return slog.GroupValue(
slog.String("name", cr.Name),
slog.String("rule", string(cr.Rule)))
}
func cr(name string, rule config.Rule) CheckResult {
return CheckResult{
Name: name,
Rule: rule,
}
}

34
lib/http.go Normal file
View File

@@ -0,0 +1,34 @@
package lib
import (
"net/http"
"time"
)
func (s *Server) ClearCookie(w http.ResponseWriter) {
http.SetCookie(w, &http.Cookie{
Name: s.opts.CookieName,
Value: "",
Expires: time.Now().Add(-1 * time.Hour),
MaxAge: -1,
Domain: s.opts.CookieDomain,
Partitioned: s.opts.CookiePartitioned,
SameSite: http.SameSiteLaxMode,
})
}
// https://github.com/oauth2-proxy/oauth2-proxy/blob/master/pkg/upstream/http.go#L124
type UnixRoundTripper struct {
Transport *http.Transport
}
// set bare minimum stuff
func (t UnixRoundTripper) RoundTrip(req *http.Request) (*http.Response, error) {
req = req.Clone(req.Context())
if req.Host == "" {
req.Host = "localhost"
}
req.URL.Host = req.Host // proxy error: no Host in request URL
req.URL.Scheme = "http" // make http.Transport happy and avoid an infinite recursion
return t.Transport.RoundTrip(req)
}

32
lib/policy/bot.go Normal file
View File

@@ -0,0 +1,32 @@
package policy
import (
"fmt"
"regexp"
"github.com/TecharoHQ/anubis/internal"
"github.com/TecharoHQ/anubis/lib/policy/config"
"github.com/yl2chen/cidranger"
)
type Bot struct {
Name string
UserAgent *regexp.Regexp
Path *regexp.Regexp
Action config.Rule `json:"action"`
Challenge *config.ChallengeRules
Ranger cidranger.Ranger
}
func (b Bot) Hash() (string, error) {
var pathRex string
if b.Path != nil {
pathRex = b.Path.String()
}
var userAgentRex string
if b.UserAgent != nil {
userAgentRex = b.UserAgent.String()
}
return internal.SHA256sum(fmt.Sprintf("%s::%s::%s", b.Name, pathRex, userAgentRex)), nil
}

View File

@@ -7,6 +7,17 @@ import (
"regexp"
)
var (
ErrNoBotRulesDefined = errors.New("config: must define at least one (1) bot rule")
ErrBotMustHaveName = errors.New("config.Bot: must set name")
ErrBotMustHaveUserAgentOrPath = errors.New("config.Bot: must set either user_agent_regex, path_regex, or remote_addresses")
ErrBotMustHaveUserAgentOrPathNotBoth = errors.New("config.Bot: must set either user_agent_regex, path_regex, and not both")
ErrUnknownAction = errors.New("config.Bot: unknown action")
ErrInvalidUserAgentRegex = errors.New("config.Bot: invalid user agent regex")
ErrInvalidPathRegex = errors.New("config.Bot: invalid path regex")
ErrInvalidCIDR = errors.New("config.Bot: invalid CIDR")
)
type Rule string
const (
@@ -24,7 +35,7 @@ const (
AlgorithmSlow Algorithm = "slow"
)
type Bot struct {
type BotConfig struct {
Name string `json:"name"`
UserAgentRegex *string `json:"user_agent_regex"`
PathRegex *string `json:"path_regex"`
@@ -33,25 +44,14 @@ type Bot struct {
Challenge *ChallengeRules `json:"challenge,omitempty"`
}
var (
ErrNoBotRulesDefined = errors.New("config: must define at least one (1) bot rule")
ErrBotMustHaveName = errors.New("config.Bot: must set name")
ErrBotMustHaveUserAgentOrPath = errors.New("config.Bot: must set either user_agent_regex, path_regex, or remote_addresses")
ErrBotMustHaveUserAgentOrPathNotBoth = errors.New("config.Bot: must set either user_agent_regex, path_regex, and not both")
ErrUnknownAction = errors.New("config.Bot: unknown action")
ErrInvalidUserAgentRegex = errors.New("config.Bot: invalid user agent regex")
ErrInvalidPathRegex = errors.New("config.Bot: invalid path regex")
ErrInvalidCIDR = errors.New("config.Bot: invalid CIDR")
)
func (b Bot) Valid() error {
func (b BotConfig) Valid() error {
var errs []error
if b.Name == "" {
errs = append(errs, ErrBotMustHaveName)
}
if b.UserAgentRegex == nil && b.PathRegex == nil && (b.RemoteAddr == nil || len(b.RemoteAddr) == 0) {
if b.UserAgentRegex == nil && b.PathRegex == nil && len(b.RemoteAddr) == 0 {
errs = append(errs, ErrBotMustHaveUserAgentOrPath)
}
@@ -71,7 +71,7 @@ func (b Bot) Valid() error {
}
}
if b.RemoteAddr != nil && len(b.RemoteAddr) > 0 {
if len(b.RemoteAddr) > 0 {
for _, cidr := range b.RemoteAddr {
if _, _, err := net.ParseCIDR(cidr); err != nil {
errs = append(errs, ErrInvalidCIDR, err)
@@ -137,8 +137,8 @@ func (cr ChallengeRules) Valid() error {
}
type Config struct {
Bots []Bot `json:"bots"`
DNSBL bool `json:"dnsbl"`
Bots []BotConfig `json:"bots"`
DNSBL bool `json:"dnsbl"`
}
func (c Config) Valid() error {

View File

@@ -13,12 +13,12 @@ func p[V any](v V) *V { return &v }
func TestBotValid(t *testing.T) {
var tests = []struct {
name string
bot Bot
bot BotConfig
err error
}{
{
name: "simple user agent",
bot: Bot{
bot: BotConfig{
Name: "mozilla-ua",
Action: RuleChallenge,
UserAgentRegex: p("Mozilla"),
@@ -27,7 +27,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "simple path",
bot: Bot{
bot: BotConfig{
Name: "well-known-path",
Action: RuleAllow,
PathRegex: p("^/.well-known/.*$"),
@@ -36,7 +36,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "no rule name",
bot: Bot{
bot: BotConfig{
Action: RuleChallenge,
UserAgentRegex: p("Mozilla"),
},
@@ -44,7 +44,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "no rule matcher",
bot: Bot{
bot: BotConfig{
Name: "broken-rule",
Action: RuleAllow,
},
@@ -52,7 +52,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "both user-agent and path",
bot: Bot{
bot: BotConfig{
Name: "path-and-user-agent",
Action: RuleDeny,
UserAgentRegex: p("Mozilla"),
@@ -62,7 +62,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "unknown action",
bot: Bot{
bot: BotConfig{
Name: "Unknown action",
Action: RuleUnknown,
UserAgentRegex: p("Mozilla"),
@@ -71,7 +71,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "invalid user agent regex",
bot: Bot{
bot: BotConfig{
Name: "mozilla-ua",
Action: RuleChallenge,
UserAgentRegex: p("a(b"),
@@ -80,7 +80,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "invalid path regex",
bot: Bot{
bot: BotConfig{
Name: "mozilla-ua",
Action: RuleChallenge,
PathRegex: p("a(b"),
@@ -89,7 +89,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "challenge difficulty too low",
bot: Bot{
bot: BotConfig{
Name: "mozilla-ua",
Action: RuleChallenge,
PathRegex: p("Mozilla"),
@@ -103,7 +103,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "challenge difficulty too high",
bot: Bot{
bot: BotConfig{
Name: "mozilla-ua",
Action: RuleChallenge,
PathRegex: p("Mozilla"),
@@ -117,7 +117,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "challenge wrong algorithm",
bot: Bot{
bot: BotConfig{
Name: "mozilla-ua",
Action: RuleChallenge,
PathRegex: p("Mozilla"),
@@ -131,7 +131,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "invalid cidr range",
bot: Bot{
bot: BotConfig{
Name: "mozilla-ua",
Action: RuleAllow,
RemoteAddr: []string{"0.0.0.0/33"},
@@ -140,7 +140,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "only filter by IP range",
bot: Bot{
bot: BotConfig{
Name: "mozilla-ua",
Action: RuleAllow,
RemoteAddr: []string{"0.0.0.0/0"},
@@ -149,7 +149,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "filter by user agent and IP range",
bot: Bot{
bot: BotConfig{
Name: "mozilla-ua",
Action: RuleAllow,
UserAgentRegex: p("Mozilla"),
@@ -159,7 +159,7 @@ func TestBotValid(t *testing.T) {
},
{
name: "filter by path and IP range",
bot: Bot{
bot: BotConfig{
Name: "mozilla-ua",
Action: RuleAllow,
PathRegex: p("^.*$"),

121
lib/policy/policy.go Normal file
View File

@@ -0,0 +1,121 @@
package policy
import (
"encoding/json"
"errors"
"fmt"
"io"
"net"
"regexp"
"github.com/prometheus/client_golang/prometheus"
"github.com/prometheus/client_golang/prometheus/promauto"
"github.com/yl2chen/cidranger"
"github.com/TecharoHQ/anubis/lib/policy/config"
)
var (
PolicyApplications = promauto.NewCounterVec(prometheus.CounterOpts{
Name: "anubis_policy_results",
Help: "The results of each policy rule",
}, []string{"rule", "action"})
)
type ParsedConfig struct {
orig config.Config
Bots []Bot
DNSBL bool
DefaultDifficulty int
}
func NewParsedConfig(orig config.Config) *ParsedConfig {
return &ParsedConfig{
orig: orig,
}
}
func ParseConfig(fin io.Reader, fname string, defaultDifficulty int) (*ParsedConfig, error) {
var c config.Config
if err := json.NewDecoder(fin).Decode(&c); err != nil {
return nil, fmt.Errorf("can't parse policy config JSON %s: %w", fname, err)
}
if err := c.Valid(); err != nil {
return nil, err
}
var validationErrs []error
result := NewParsedConfig(c)
result.DefaultDifficulty = defaultDifficulty
for _, b := range c.Bots {
if berr := b.Valid(); berr != nil {
validationErrs = append(validationErrs, berr)
continue
}
parsedBot := Bot{
Name: b.Name,
Action: b.Action,
}
if len(b.RemoteAddr) > 0 {
parsedBot.Ranger = cidranger.NewPCTrieRanger()
for _, cidr := range b.RemoteAddr {
_, rng, err := net.ParseCIDR(cidr)
if err != nil {
return nil, fmt.Errorf("[unexpected] range %s not parsing: %w", cidr, err)
}
parsedBot.Ranger.Insert(cidranger.NewBasicRangerEntry(*rng))
}
}
if b.UserAgentRegex != nil {
userAgent, err := regexp.Compile(*b.UserAgentRegex)
if err != nil {
validationErrs = append(validationErrs, fmt.Errorf("while compiling user agent regexp: %w", err))
continue
} else {
parsedBot.UserAgent = userAgent
}
}
if b.PathRegex != nil {
path, err := regexp.Compile(*b.PathRegex)
if err != nil {
validationErrs = append(validationErrs, fmt.Errorf("while compiling path regexp: %w", err))
continue
} else {
parsedBot.Path = path
}
}
if b.Challenge == nil {
parsedBot.Challenge = &config.ChallengeRules{
Difficulty: defaultDifficulty,
ReportAs: defaultDifficulty,
Algorithm: config.AlgorithmFast,
}
} else {
parsedBot.Challenge = b.Challenge
if parsedBot.Challenge.Algorithm == config.AlgorithmUnknown {
parsedBot.Challenge.Algorithm = config.AlgorithmFast
}
}
result.Bots = append(result.Bots, parsedBot)
}
if len(validationErrs) > 0 {
return nil, fmt.Errorf("errors validating policy config JSON %s: %w", fname, errors.Join(validationErrs...))
}
result.DNSBL = c.DNSBL
return result, nil
}

View File

@@ -1,25 +1,28 @@
package main
package policy
import (
"os"
"path/filepath"
"testing"
"github.com/TecharoHQ/anubis"
"github.com/TecharoHQ/anubis/data"
)
func TestDefaultPolicyMustParse(t *testing.T) {
fin, err := static.Open("botPolicies.json")
fin, err := data.BotPolicies.Open("botPolicies.json")
if err != nil {
t.Fatal(err)
}
defer fin.Close()
if _, err := parseConfig(fin, "botPolicies.json", defaultDifficulty); err != nil {
if _, err := ParseConfig(fin, "botPolicies.json", anubis.DefaultDifficulty); err != nil {
t.Fatalf("can't parse config: %v", err)
}
}
func TestGoodConfigs(t *testing.T) {
finfos, err := os.ReadDir("internal/config/testdata/good")
finfos, err := os.ReadDir("config/testdata/good")
if err != nil {
t.Fatal(err)
}
@@ -27,13 +30,13 @@ func TestGoodConfigs(t *testing.T) {
for _, st := range finfos {
st := st
t.Run(st.Name(), func(t *testing.T) {
fin, err := os.Open(filepath.Join("internal", "config", "testdata", "good", st.Name()))
fin, err := os.Open(filepath.Join("config", "testdata", "good", st.Name()))
if err != nil {
t.Fatal(err)
}
defer fin.Close()
if _, err := parseConfig(fin, fin.Name(), defaultDifficulty); err != nil {
if _, err := ParseConfig(fin, fin.Name(), anubis.DefaultDifficulty); err != nil {
t.Fatal(err)
}
})
@@ -41,7 +44,7 @@ func TestGoodConfigs(t *testing.T) {
}
func TestBadConfigs(t *testing.T) {
finfos, err := os.ReadDir("internal/config/testdata/bad")
finfos, err := os.ReadDir("config/testdata/bad")
if err != nil {
t.Fatal(err)
}
@@ -49,13 +52,13 @@ func TestBadConfigs(t *testing.T) {
for _, st := range finfos {
st := st
t.Run(st.Name(), func(t *testing.T) {
fin, err := os.Open(filepath.Join("internal", "config", "testdata", "bad", st.Name()))
fin, err := os.Open(filepath.Join("config", "testdata", "bad", st.Name()))
if err != nil {
t.Fatal(err)
}
defer fin.Close()
if _, err := parseConfig(fin, fin.Name(), defaultDifficulty); err == nil {
if _, err := ParseConfig(fin, fin.Name(), anubis.DefaultDifficulty); err == nil {
t.Fatal(err)
} else {
t.Log(err)

9
lib/random.go Normal file
View File

@@ -0,0 +1,9 @@
package lib
import (
"math/rand"
)
func randomJitter() bool {
return rand.Intn(100) > 10
}

14
web/embed.go Normal file
View File

@@ -0,0 +1,14 @@
package web
import "embed"
//go:generate go tool github.com/a-h/templ/cmd/templ generate
//go:generate esbuild js/main.mjs --sourcemap --bundle --minify --outfile=static/js/main.mjs
//go:generate gzip -f -k static/js/main.mjs
//go:generate zstd -f -k --ultra -22 static/js/main.mjs
//go:generate brotli -fZk static/js/main.mjs
var (
//go:embed static
Static embed.FS
)

15
web/index.go Normal file
View File

@@ -0,0 +1,15 @@
package web
import "github.com/a-h/templ"
func Base(title string, body templ.Component) templ.Component {
return base(title, body)
}
func Index() templ.Component {
return index()
}
func ErrorPage(msg string) templ.Component {
return errorPage(msg)
}

View File

@@ -1,4 +1,4 @@
package main
package web
import (
"github.com/TecharoHQ/anubis"

View File

@@ -1,7 +1,7 @@
// Code generated by templ - DO NOT EDIT.
// templ: version: v0.3.833
package main
package web
//lint:file-ignore SA4006 This context is only used if a nested component is present.

View File

Before

Width:  |  Height:  |  Size: 58 KiB

After

Width:  |  Height:  |  Size: 58 KiB

View File

Before

Width:  |  Height:  |  Size: 48 KiB

After

Width:  |  Height:  |  Size: 48 KiB

View File

Before

Width:  |  Height:  |  Size: 50 KiB

After

Width:  |  Height:  |  Size: 50 KiB

View File

@@ -1,6 +1,6 @@
// Code generated by templ - DO NOT EDIT.
// templ: version: v0.3.833
// templ: version: v0.3.850
package xess
//lint:file-ignore SA4006 This context is only used if a nested component is present.