Structured materials for technical reviewers, security evaluators, and product teams. The full evaluation path takes approximately 15 minutes.
The browser vault is running at secure.imagineqira.com/vault. It seals and opens Vault V2 packages entirely in the tab — no account, no server. The earlier policy/envelope demo dashboard is not currently hosted here; the capture gateway it previewed is a pilot, described at /platform/.
Open VaultA concise memo covering the product overview, strongest capabilities, trust model, hardening status, limitations, and key files for code review. Written for fast comprehension.
Technical specification of every trust-related field in bundle verification output. Covers packet vs. envelope layers, integrity vs. signature vs. trust, the full truth table for all verification states, and what the semantics do not cover.
Copy-paste commands for running tests, starting services, building envelopes, exporting signed bundles, verifying with correct and incorrect keys, and confirming provider lockdown restrictions. Every command is tested and minimal.
Step-by-step walkthrough of the strongest end-to-end workflow: build an envelope, compare policies, browse saved artifacts, export a signed bundle, verify it, then demonstrate trust failure states.
Current trust boundaries, adversary classes, out-of-scope claims, known weak areas, and findings from the internal adversarial review. Honest about what is protected and what is not.
If you have the repository locally, run the preparation script to generate a saved envelope, a signed bundle, and a verification summary:
uv run python scripts/prepare_review_artifacts.py
Artifacts land under data/review/, data/runs/, and data/bundles/. All secrets used are printed in the output. You can also run the full workflow interactively through the browser vault without any local setup.
Does the system hold together? Are the trust semantics clear? Does the verification output make sense? Are the protocol boundaries well-defined?
Are the hardening measures adequate for the stated trust model? Are the documented limitations accurate? What would you harden next?
Is it immediately clear what this system does and does not do? Does the positioning feel honest? Is anything confusing or misleading?
What would prevent you from using or recommending this system? What is missing? What feels weak?
Start with the live demo or contact us for a guided technical session.