Files
splunk-security_content/detections/cloud/github_commit_in_develop.yml
T
2024-09-30 18:10:16 +05:30

53 lines
3.1 KiB
YAML

name: Github Commit In Develop
id: f3030cb6-0b02-11ec-8f22-acde48001122
version: 3
date: '2024-09-30'
author: Teoderick Contreras, Splunk
status: production
type: Anomaly
description: The following analytic detects commits pushed directly to the 'develop' or 'main' branches in a GitHub repository. It leverages GitHub logs, focusing on commit metadata such as author details, commit messages, and timestamps. This activity is significant as direct commits to these branches can bypass the review process, potentially introducing unvetted changes. If confirmed malicious, this could lead to unauthorized code modifications, introducing vulnerabilities or backdoors into the codebase, and compromising the integrity of the development lifecycle.
data_source:
- GitHub
search: '`github` branches{}.name = main OR branches{}.name = develop | stats count min(_time) as firstTime max(_time) as lastTime by commit.author.html_url commit.commit.author.email commit.author.login commit.commit.message repository.pushed_at commit.commit.committer.date | eval phase="code" | `security_content_ctime(firstTime)` | `security_content_ctime(lastTime)` | `github_commit_in_develop_filter`'
how_to_implement: To successfully implement this search, you need to be ingesting logs related to github logs having the fork, commit, push metadata that can be use to monitor the changes in a github project.
known_false_positives: admin can do changes directly to develop branch
references:
- https://www.redhat.com/en/topics/devops/what-is-devsecops
drilldown_searches:
- name: View the detection results for $commit.commit.author.email$
search: '%orginal_detection_search% | search commit.commit.author.email = $commit.commit.author.email$'
earliest_offset: $info_min_time$
latest_offset: $info_max_time$
- name: View risk events for the last 7 days for $commit.commit.author.email$
search: '| from datamodel Risk.All_Risk | search normalized_risk_object IN ($commit.commit.author.email$) starthoursago=168 endhoursago=1 | stats count min(_time) as firstTime max(_time) as lastTime values(search_name) as "Search Name" values(risk_message) as "Risk Message" values(analyticstories) as "Analytic Stories" values(annotations._all) as "Annotations" values(annotations.mitre_attack.mitre_tactic) as "ATT&CK Tactics" by normalized_risk_object | `security_content_ctime(firstTime)` | `security_content_ctime(lastTime)`'
earliest_offset: $info_min_time$
latest_offset: $info_max_time$
tags:
analytic_story:
- Dev Sec Ops
asset_type: GitHub
confidence: 30
impact: 30
message: Suspicious commit by $commit.commit.author.email$ to develop branch
mitre_attack_id:
- T1199
observable:
- name: commit.commit.author.email
type: User
role:
- Victim
product:
- Splunk Enterprise
- Splunk Enterprise Security
- Splunk Cloud
required_fields:
- _time
risk_score: 9
security_domain: endpoint
tests:
- name: True Positive Test
attack_data:
- data: https://media.githubusercontent.com/media/splunk/attack_data/master/datasets/attack_techniques/T1199/github_push_master/github_push_develop.json
source: github
sourcetype: aws:firehose:json