Repository navigation
Security and Trust
No. Not with real secrets. There is no implemented cryptography, no released software, and no audit. Every security property in the documentation is a design intention, not a verified guarantee. The always-current answer lives in the Trust Center.
No perfect anonymity. No unbreakable encryption. No forensic erasure. No spyware-proof screens. No impossible-to-stop communication. These phrases are review-blockers in this project — honest language is enforced, not just encouraged.
- Compromised devices — malware with sufficient privilege sees what you see.
- Malicious room members — encryption cannot protect you from people you invited.
- Global traffic analysis — a sufficiently positioned observer can correlate timing and size.
- Cameras pointed at screens — physics wins.
Full detail: Threat Model.
- CI on every change: typecheck, lint, tests, build, four privacy guards (no telemetry, no external assets, import boundaries, no forbidden storage APIs) and doc-link integrity.
- CodeQL static analysis (security-extended), dependency review, Dependabot, secret scanning + push protection.
- Branch protection: PRs with required checks and review; workflow actions pinned by commit SHA.
- A public PBOM (Privacy Bill of Materials): anything the software does that isn't listed there is treated as a bug.
Privately, please: Security tab → "Report a vulnerability" on the repository. Policy: SECURITY.md.
FreeLayer — foundation stage · not ready for real secrets · code AGPL-3.0-or-later, docs CC BY-SA 4.0 · canonical docs live in docs/
FreeLayer Wiki