MeshKit
Security

Security Reviews

Evidence pages for security review work.

Security reviews are evidence, not onboarding material. Use this page when you need to plan, run, or interpret a MeshKit security review.

Adopters should start with Security and Threat model. Maintainers and reviewers should use this page to keep review scope repeatable.

Review Areas

AreaQuestions
Envelope encryptionAre authenticated fields covered? Are algorithms and versions explicit?
CID verificationAre wrong bytes rejected before decryption?
Identity lifecycleAre export, import, publish, resolve, recovery, and logs safe?
Sharing and revocationDo expired/revoked/policy-denied shares fail closed?
Metadata leakageAre sensitive relationships minimized and documented?
Provider boundariesAre bytes, metadata, policy, persistence, and logs separated correctly?
Runtime key storageDo mobile/web claims match real bridge capabilities?
TelemetryAre secrets and plaintext redacted from events and support bundles?
Supply chainAre release scripts, package names, and publishing tokens controlled?

Evidence To Capture

  • Test or validation command
  • Environment summary with secret values redacted
  • Provider configuration shape
  • MeshKitError.code for expected failures
  • Proof summaries, not plaintext
  • Capsule, policy, vault, or capability IDs when relevant
  • Link to source commit or release version

What Not To Capture

  • private keys
  • bearer tokens
  • provider response bodies containing secrets
  • decrypted user content
  • capability token secrets
  • unredacted .npmrc or CI secret output

Review Output Template

Scope:
Version:
Provider boundary:
Runtime:
Workflows reviewed:
Commands run:
Expected failures:
Findings:
Residual risks:
Follow-up owner:

Next Steps

On this page