Scope & Rules of Engagement
In-scope — web / API
| Asset | Notes |
|---|---|
corezoid.com |
Marketing / landing |
corezoid.com/api |
Primary API |
superadmin.corezoid.com |
Operator/superadmin panel |
admin.corezoid.com |
Tenant admin UI (HAR file provided) |
account.corezoid.com |
Auth / SSO / account management |
jira.corezoid.com |
Atlassian Jira instance |
openapi.corezoid.com |
OpenAPI specification host |
*.corezoid.com |
All discovered subdomains (enumerated in
recon/subdomains.md) |
simulator.company |
Simulator product root |
doc.simulator.company |
Documentation site |
mw.simulator.company |
Simulator middleware (HAR file provided) |
sim.simulator.company |
Simulator UI / app |
vpn.corezoid.com |
VPN endpoint — fingerprint-only, no auth brute |
In-scope — source code
All repositories under github.com/corezoid/ are in scope
for cloning, reading, SAST, and secret scanning. Explicitly named:
apigw— API gatewayhelm— Helm chartsdbcall— DB connectorcorezoid_ansible_roles— Ansible rolesaccount— Account servicegitcall— Git connectorterraform— Infrastructure-as-codeobservability— Monitoring stackgitcall-livekit-agent— LiveKit integrationcorezoid-ai-doc— AI documentation (reference material)- Any other public repo discovered under the org.
Out of scope
- Third-party SaaS and partner systems (Atlassian Cloud, AWS console, GitHub.com itself, CI providers, etc.)
- Customer data belonging to Corezoid tenants — do not attempt to pivot into customer workflows
- Payment processors and any PCI-scoped systems
- Physical and social-engineering attacks
- DoS / resource-exhaustion / fork-bomb payloads
- Any asset not explicitly listed here or discovered under
*.corezoid.com/*.simulator.companyduring recon
Testing rules (hard limits)
- Non-destructive only. No
DROP, noDELETE *, no cache flushes, no mass account creation. Proof-of-concept for writes must use a single tester-controlled record and be cleaned up. - No production data exfiltration. If a vulnerability exposes customer PII, capture the proof (one row, redacted), not the dataset.
- Rate limits — CONSERVATIVE mode (CTO directive
2026-04-26). Max 10 req/s per host across all
tools. Nuclei: default tags only, no
-severity critical -t intrusive, no-tags dos,brute,fuzz. Slow down aggressively if WAF/RL fires. - Auth brute-forcing limited. No password spray.
Check weak default credentials (
admin:admin, etc.) only against test/staging endpoints that are clearly marked. - PoC depth. For RCE, SSRF, SQLi — prove it with a
low-impact canary (read
/etc/hostname, resolve a DNS callback,SELECT @@version), not full exploitation. - Log everything. Every request that resulted in a
finding lands in
evidence/CRZ-NNN-*. - Stop if in doubt. Any ambiguous blast-radius decision → pause and ask the CTO.
- Phase-gated execution (CTO directive 2026-04-26). Each of the 9 phases completes with a summary, then pause for CTO confirmation before starting the next phase. Do not run phases back-to-back autonomously.
Authentication material (authorized)
Provided by the CTO at engagement start:
admin.corezoid.com.har+admin.corezoid.com_cookies.txt— admin UI sessionmw.simulator.company.har+mw.simulator.company_cookies.txt— simulator MW session
These are live authenticated sessions. Using them to test authenticated endpoints is authorized. Do not share, publish, or commit these files to any remote repository.
Reporting classification
Severity follows CVSS 3.1-ish informal bands:
- Critical (9.0+): RCE, auth bypass to admin, mass data exposure, secret leakage with live credentials
- High (7.0–8.9): Privilege escalation, stored XSS in admin, SSRF to internal metadata, IDOR exposing other tenants
- Medium (4.0–6.9): Reflected XSS, CSRF on sensitive actions, weak TLS, missing security headers on sensitive endpoints
- Low (<4.0): Information disclosure (versions, stack traces), minor misconfig, missing headers on low-value pages
- Info: Hardening recommendations, observations