snare
is a GitHub webhooks daemon. When snare
receives a webhook event from
a given repository, it authenticates the request, and then executes a
user-defined "per-repo program" with information about the webhook event.
snare
requires rustc-1.40.0 or greater.
To install snare
on a per-user basis, use cargo install snare
.
To install snare
globally and/or for packaging purposes, download the latest
stable version from snare
's homepage.
You may use cargo
to build snare locally or you may use the Makefile
to
build and install snare
in traditional Unix fashion. make install
defaults
to installing in /usr/local
: you can override this by setting the PREFIX
variable to another path (e.g. PREFIX=/opt/local make
).
snare
has the following command-line format:
Usage: snare [-c <config-path>] [-d]
where:
-c <config-path>
is a path to asnare.conf
configuration file. If not specified,snare
will assume the configuration file is located at/etc/snare/snare.conf
.-d
tellssnare
not to daemonise: in other words,snare
stays in the foreground. This can be useful for debugging.
The man page for snare contains more details.
The minimal recommended configuration file is:
listen = "<ip-address>:<port>";
github {
match ".*" {
cmd = "/path/to/prps/%o/%r %e %j";
errorcmd = "cat %s | mailx -s \"snare error: github.com/%o/%r: %x %?\" someone@example.com";
secret = "<secret>";
}
}
where:
-
ip-address
is either an IPv4 or IPv6 address andport
a port on which an HTTP server will listen. -
cmd
is the command that will be executed when a webhook is received. In this case,/path/to/prps
is a path to a directory where per-repo programs are stored. For a repositoryrepo
owned byowner
the command:/path/to/prps/<owner>/<repo> <event> <path-to-github-json>
will be run. The file
<repo>
must be executable. Note that commands are run with their current working directory set to a temporary directory to which they can freely write and which will be automatically removed when they have completed. -
errorcmd
is the command that will be run when acmd
exits unsuccessfully. In this example, an email is sent tosomeone@example.com
with a title indicating the reason for non-zero exit (%x %?
) and a body consisting of the comined stedrr/stdout. This assumes that you have installed, set-up, and enabled a suitablesendmail
clone. -
secret
is the GitHub secret used to sign the webhook request and thus allowingsnare
to tell the difference between genuine webhook requests and those from malfeasants.
The man page for snare.conf contains the complete list of configuration options.
snare
can be used to run any command runnable from the Unix shell. The
"per-repo program" model as documented above is one common way of doing this.
For example, snare
's GitHub
repository is
https://github.com/softdevteam/snare
.
If we set up a web hook up for that repository that notifies us of pull request
events, then with the above snare.conf
, the command:
/path/to/prps/softdevteam/snare pull_request /path/to/json
will be executed, where: pull_request
is the name of the GitHub event; and
/path/to/json
is a path to a file containing the complete GitHub JSON for
that event. The softdevteam/snare
program can then execute whatever it wants.
In order to work out precisely what event has happened, you will need to read
GitHub's webhooks documentation.
Users can write per-repo programs in whatever system/language they wish, so
long as the matching file is marked as executable. The following simple example
uses shell script to send a list of commits and diffs to the address specified
in $EMAIL
on each push
to master. It works for any public GitHub
repository:
#! /bin/sh
set -eufx
# A list of email addresses separated by spaces.
EMAILS="someone@example.com someone.else@example.com"
# A GitHub URL either https or git.
REPO_URL="git@github.com:owner/repo.git"
if [ "$1" != "push" ]; then
exit 0
fi
ref=`jq .ref "$2" | tr -d '\"'`
if [ "$ref" != "refs/heads/master" ]; then
exit 0
fi
repo_fullname=`jq .repository.full_name "$2" | tr -d '\"'`
repo_url=`jq .repository.html_url "$2" | tr -d '\"'`
before_hash=`jq .before "$2" | tr -d '\"'`
after_hash=`jq .after "$2" | tr -d '\"'`
echo "$before_hash" | grep -E "^[a-fA-F0-9]+$" 2>&1 > /dev/null
echo "$after_hash" | grep -E "^[a-fA-F0-9]+$" 2>&1 > /dev/null
git clone "$REPO_URL" repo
cd repo
for email in `echo "$EMAILS"`; do
git log --reverse -p "$before_hash..$after_hash" | mail -s "Push to $repo_fullname" "$email"
done
where jq
is a command-line JSON processor.
Depending on your needs, you can make this type of script arbitrarily more
complex and powerful (e.g. not cloning afresh on each pull).
Note that this program is deliberately untrusting of external input: it is
careful to quote all arguments obtained from JSON; and it uses a fixed
directory name (repo
) rather than use a file name from JSON that might
include characters (e.g. ../..
) that would cause the script to leak data
about other parts of the file system.
snare
runs an HTTP server which GitHub can send webhook requests to.
Configuring a webhook for a given GitHub repository is relatively simple: go to
that repository, then Settings > Webhooks > Add webhook
. For payload
,
specify http://yourmachine.com:port/
, specify a secret
(which you will then
reuse as the secret
in snare.conf
) and then choose which events you wish
GitHub to deliver. For example, the default Just the push event
works well
with the email diff sending per-repo program above, but you can specify
whichever events you wish.
Once you have set up your webhook, GitHub will automatically send a test "ping"
message. snare
acknowledges pings (i.e. confirms to GitHub that it is
running), but does not execute any commands in response to pings.
snare
runs an HTTP server. If you wish, as is recommended, to send your
webhooks over an encrypted connection, you will need to run a proxy in front of
snare e.g.
nginx or
relayd.