| Version | Supported |
|---|---|
main (latest) |
Yes |
Do not open a public GitHub issue for security vulnerabilities.
<<<<<<< HEAD Use GitHub Private Vulnerability Reporting on this repository.
Include:
- Affected endpoints, events, or components
- Reproduction steps and expected vs actual behavior
- Impact (confidentiality / integrity / availability)
- PoC only when it stays non-destructive
- Acknowledgement when maintainers triage the report
- Coordination on fix and disclosure timing ======= Use GitHub Private Vulnerability Reporting.
Include reproduction steps, affected UI flows, and impact. Prefer non-destructive PoCs.
- Triage acknowledgement
- Coordinated fix and disclosure
9d2bd3842e9cc45c912ac09908bd70f02464070f
- Credit when appropriate and desired
<<<<<<< HEAD
- Auth bypass, JWT misuse, rate-limit evasion with security impact
- IDOR / privilege escalation
- Injection (NoSQL, path traversal, unsafe deserialization)
- E2E / X5 regressions (server storing or notifying plaintext chat content)
- Broken sealed call signaling, story envelopes, or push blindness
- Secrets leakage in logs, artifacts, or CI
- Lost client keyrings /
keys.txt - Compromised user devices
- Pure DoS without a practical, actionable exploit
- Issues solely in Frontend or QuantumAI (report to those repos)
- Helmet CSP for API responses (
default-src 'none'),nosniff,frameguard: deny - Avatar / group photo / story image uploads allowlist JPEG, PNG, WebP, GIF only (no SVG)
- Served media uses allowlisted
Content-Type(does not trust client MIME for raster images)
Good-faith research that follows this policy and avoids privacy abuse or destructive testing will not be pursued legally by the maintainers.
- Private key / keyring exposure (logs, exports, storage bugs)
- Plaintext message or call-signaling leakage to the network
- XSS or client bugs that steal session tokens or key material
- Broken seal/unseal or key-import validation
- Dependency issues with a realistic exploit path in this app
- Users who lose
keys.txtor clear site data - Compromised devices / malicious browser extensions (general)
- Backend-only issues (report to QuantumChat-Backend)
- Content-Security-Policy on the SPA (
script-src 'self', no inline scripts) - Chat text rendered as React text nodes (escaped)
- SVG attachments are not image-previewed; PDF iframes use
sandbox="allow-same-origin" - Avatar file picker limited to raster image types
Good-faith research that follows this policy and avoids abusing real users’ data will not be pursued legally by the maintainers.
9d2bd3842e9cc45c912ac09908bd70f02464070f