Version: 1.0 Last Updated: 2026-01-28
We actively support and provide security updates for the following versions of Thread:
| Version | Supported | End of Support |
|---|---|---|
| 0.1.x | ✅ | TBD |
| < 0.1 | ❌ | Immediately |
Support Policy:
- Latest minor version receives security patches
- Previous minor version receives critical security patches for 3 months after new release
- Major versions receive security support for 12 months after new major release
We take security vulnerabilities seriously and appreciate responsible disclosure.
DO NOT create public GitHub issues for security vulnerabilities.
Instead, please report security issues to:
Email: security@knit.li
Include in your report:
- Description of the vulnerability
- Steps to reproduce the issue
- Potential impact and severity assessment
- Suggested remediation (if available)
- Your contact information for follow-up
- Acknowledgment: Within 24 hours of submission
- Initial Assessment: Within 72 hours
- Status Update: Weekly updates on progress
- Resolution Timeline:
- Critical: 7 days
- High: 14 days
- Medium: 30 days
- Low: 90 days
- Coordinated Disclosure: We follow a 90-day disclosure timeline
- Security Advisory: Published on GitHub Security Advisories
- CVE Assignment: Requested for critical and high severity issues
- Credit: Security researchers will be credited unless they prefer to remain anonymous
Static Analysis:
- Automated scanning with Semgrep SAST
- Clippy linting with security rules
- Regular security audits
Dependency Management:
- Daily automated vulnerability scanning with
cargo-audit - Dependency review on all pull requests
- License compliance checks
- Supply chain security with
cargo-deny
Code Review:
- All code changes require review
- Security-sensitive changes require security team review
- Automated checks must pass before merge
CI/CD Security:
- Signed releases with checksums
- Reproducible builds
- Minimal build dependencies
- Isolated build environments
Artifact Verification:
# Verify release checksum
sha256sum -c thread-0.1.0-checksums.txt
# Verify with GPG (when available)
gpg --verify thread-0.1.0.tar.gz.sigSandboxing:
- Minimal required permissions
- Process isolation where applicable
- Secure defaults for all features
Data Protection:
- No credentials stored in logs
- Secure credential handling
- Encrypted data transmission
Access Control:
- Multi-factor authentication required
- Least privilege access model
- Regular access reviews
Secrets Management:
- Environment-based secrets
- No secrets in version control
- Regular secret rotation
Installation:
# Verify download authenticity
curl -LO https://github.com/knitli/thread/releases/latest/download/thread-0.1.0-x86_64-unknown-linux-gnu.tar.gz
sha256sum thread-0.1.0-x86_64-unknown-linux-gnu.tar.gz
# Install from trusted sources only
cargo install thread-flow # From crates.io
# or
brew install knitli/tap/thread # From official tapConfiguration:
- Use environment variables for sensitive configuration
- Never commit credentials to version control
- Rotate database credentials regularly
- Use read-only database users where possible
Network Security:
- Use TLS for all database connections
- Enable SSL mode for PostgreSQL:
?sslmode=require - Implement firewall rules for database access
- Use private networks for database connections
Development Security:
-
Run security checks before committing:
cargo audit cargo clippy -- -D warnings
-
Never commit:
- API keys or credentials
- Private keys or certificates
- Database connection strings with passwords
.envfiles with secrets
-
Use pre-commit hooks:
hk install # Install git hooks
Dependency Updates:
- Review
cargo updatechanges carefully - Check for security advisories before updating
- Test thoroughly after dependency updates
PostgreSQL:
- Use connection pooling with reasonable limits
- Implement query timeouts
- Sanitize user input (handled by sqlx)
- Use prepared statements (default with sqlx)
D1 (Cloudflare):
- Rate limiting applied automatically
- Row limits enforced
- Sandboxed execution environment
WASM Sandboxing:
- Limited system access
- No filesystem access
- Memory limits enforced
- CPU time limits
Cloudflare Workers Security:
- Isolated V8 contexts
- Automatic DDoS protection
- Built-in rate limiting
- Secure execution environment
System Access:
- File system access as configured
- Network access as configured
- Runs with user permissions
- Systemd service isolation (recommended)
Currently no active security advisories.
None at this time.
- GitHub: Watch repository → Custom → Security alerts
- Email: Subscribe to security@knit.li mailing list
- RSS: https://github.com/knitli/thread/security/advisories.atom
| Severity | Response Time | Patch Release | Communication |
|---|---|---|---|
| Critical | 24 hours | 7 days | Immediate advisory |
| High | 48 hours | 14 days | Security advisory |
| Medium | 1 week | 30 days | Release notes |
| Low | 2 weeks | 90 days | Release notes |
Severity Criteria:
- Critical: Remote code execution, privilege escalation, data breach
- High: Authentication bypass, significant data exposure, DoS
- Medium: Information disclosure, limited DoS, CSRF
- Low: Minor information leaks, theoretical attacks
| Date | Type | Auditor | Report |
|---|---|---|---|
| TBD | External Security Audit | TBD | TBD |
- OWASP Top 10: Addressed in design and implementation
- CWE Top 25: Mitigated through secure coding practices
- SANS Top 25: Covered by security controls
- SOC 2: Planned for future
- ISO 27001: Planned for future
For Development:
cargo-audit- Vulnerability scanningcargo-deny- Supply chain securitycargo-outdated- Dependency updatescargo-geiger- Unsafe code detection
For Operations:
fail2ban- Intrusion preventionufw- Firewall configurationLet's Encrypt- TLS certificatesVault- Secret management
# Install security tooling
cargo install cargo-audit cargo-deny cargo-outdated cargo-geiger
# Run security checks
cargo audit
cargo deny check all
cargo geiger- Security Issues: security@knit.li
- General Questions: support@knit.li
- Bug Reports: https://github.com/knitli/thread/issues (non-security)
We would like to thank the following security researchers for responsibly disclosing vulnerabilities:
(None at this time)
Responsible Disclosure: We are committed to working with security researchers through coordinated disclosure. Thank you for helping keep Thread and our users safe.
Last Updated: 2026-01-28 Next Review: 2026-04-28 (Quarterly)