Export
Export your threat models for sharing, review, or compliance documentation.
Formats
| Format | Content | Best for |
|---|---|---|
| Report (HTML) | Full stakeholder report: model structure, controls with implementation status, assurance summary, compliance coverage, risk tiers | Sharing with stakeholders, GRC, CISOs — opens in any browser |
| Same rich report as HTML, rendered as PDF | Printing, archiving, email attachments | |
| CSV | Raw model structure (assets, attackers, COs, assumptions) | Importing into spreadsheets, data analysis, tooling |
| Audit archive (JSON) | Self-contained envelope of the model's current state: the latest version and its controls, live assertions with the CI runs behind their Tier 1/Tier 2 statuses, decisions in force, findings, and instance sufficiency signatures | Audit retention, cross-instance portability, independent signature verification |
How to export
- Go to the Models page or Assurance Dashboard
- Select a model
- Click Report (HTML), PDF, CSV, or Audit archive
The Report and PDF formats include the enriched report: trust boundaries, assets, attackers, control objectives, implementation controls with traffic-light status, assurance summary, risk tier breakdown, verification status, and compliance coverage (if a framework is selected). Whether a delegation to another model's control credits its objective depends on that model's controls, so a single model's export shows the delegation, not the provider's proof. Manual attestations are preserved.
The CSV format includes the raw model structure: trust boundaries, assets (with security properties), attackers, control objectives, and assumptions.
Audit archive (JSON)
The audit archive is a self-contained JSON envelope carrying the model's current state and the evidence an auditor needs to re-verify every current "verified implemented" claim — without needing access to the origin Mipiti instance. History (earlier versions, past changes to the controls, activity) stays on the platform; export again whenever you need a newer snapshot.
What it contains:
- The latest version of the threat model
- The controls published for it, with their current implementation status (while a control build is running, the last published controls)
- The live assertions with their full content, Tier 1 and Tier 2 statuses, attested flags, and the CI run behind each status with the attestation that signed it (Sigstore bundle for OIDC-capable CIs; workspace ECDSA signature for non-OIDC CIs; or a customer-keyed offline DSSE / in-toto bundle for air-gapped / non-Sigstore CI)
- Open findings, and the findings a person closed
- The decisions in force: risk acceptances, assumption overrides and live attestations
- The signature of each control's current sufficiency judgement (ECDSA P-256 over the canonical verdict + a content hash of every assertion that fed the decision)
Independently verifiable:
- CI Sigstore bundles verify against Sigstore's public Fulcio root + Rekor transparency log — no dependency on the origin instance, and the Rekor inclusion proof is portable for years.
- Workspace ECDSA submission signatures verify against the origin workspace's published public key.
- Customer-keyed offline DSSE bundles verify fully offline — no Fulcio, no Rekor, no network. The auditor runs
mipiti-verify audit <model_id> --expected-customer-key <fingerprint>, pinning the fingerprint received out-of-band directly from the customer (never from the bundle's convenience PEM nor Mipiti's stored copy). Trust derives from that customer-supplied pin, so this path is vendor-independent without caveat: Mipiti holds no private key and is pure transport — it can neither forge nor substitute the attestation, and a swapped key fails closed. This is the additional path for environments that structurally cannot use Sigstore (air-gapped, regulated, or non-OIDC CI); it does not replace the Sigstore path where Sigstore is reachable. See Assertion Verification for the customer setup. - Instance sufficiency signatures verify against the origin instance's public ECDSA key, which the target instance imports via the Admin Panel's Trusted Signers section.
Tampering with any signed field — flipping a Tier 1/Tier 2 verdict, mutating an assertion's parameters, adding or removing assertions after signing — invalidates the relevant signature on verify.
Importing an archive
Another Mipiti instance (or the same one) can import the archive via the Import audit archive button on the Models page. It becomes version 1 of a new model under a fresh model_id, so the same envelope can be imported any number of times. Title collisions in the target workspace auto-suffix (imported YYYY-MM-DD) — you can rename afterwards.
The archive's verdicts and signatures are shown as the origin's claims, never as judgements made here, and an import starts no judging: the imported model reads as awaiting judgement, and the import shows an estimate of judging it. Judge it from the model page when you choose to.
Via MCP, use export_threat_model_archive / import_threat_model_archive.
Viewing and verifying a sufficiency signature
On the Assurance tab, any control with an evaluated sufficiency verdict shows a small padlock next to its status badge. Click it to open the signature modal, which shows:
- The signed payload (verdict, reasoning, evaluated timestamp, sufficiency model)
- The signer fingerprint and the assertions content hash
- The raw base64 signature and the signer's public key (PEM)
- A Verify signature button that re-checks the signature server-side
- Copy bundle for offline verification with
opensslor thecryptographyPython library
Tag export
For a report that spans several models, put them in a tag and use Export on the tag's page. It downloads a signed HTML report: the delegations among the members, each with its status and whether it credits its objective, followed by every member model's full report. This is the auditor artifact for a product or an audit scope.
Jira export
For Jira integration, use the dedicated "Export to Jira" button instead. This creates structured Jira issues rather than a flat document export. Each Jira issue includes mitigation group information in the description to show which controls are alternatives and which are defense-in-depth. See the Integrations section for details.