Skip to main content

Security and Evidence FAQ

Who this is for​

This page is for workspace teams evaluating Dralvia’s security posture, evidence model, and practical use in real workflows.

Role-based start here​

  • Security lead: read the sections on explanation, maturity, and proof of value.
  • Compliance or assurance reviewer: focus on reports, EvidencePacks (Dralvia's exportable evidence reports), and audit handoff.
  • Operator: focus on what Dralvia does today and what to validate first.
  • Technical buyer: focus on the sections about AI-assisted attacks, agents, and real product scope.

Before you start​

  • Decide which workflow you are evaluating first.
  • Be clear about whether you need explanation, exportability, or action guidance.
  • Treat this FAQ as a practical product guide, not a legal policy substitute.

Step-by-step​

  1. Confirm the workflow you care about most.
  2. Review what Dralvia does on that surface today.
  3. Check whether the result is explainable and exportable enough for your team.
  4. Validate one real input before drawing conclusions.

What does Dralvia actually do today?​

Dralvia helps humans and teams decide whether they should trust, open, connect, sign, execute, or escalate risky digital interactions. Today that is strongest in phishing triage, proof/reporting, wallet approval review, and repo/package review.

Are the results explainable?​

Yes. The platform is built to show verdicts with explanations, evidence, and artifact exports instead of just a raw score.

Can we export proof for another team or an auditor?​

Yes. Reports and EvidencePacks are the main handoff format when a verdict needs to be reviewed outside the immediate operator.

How does someone verify an EvidencePack we share with them?​

Verification has two independent parts, and you do not need to understand Dralvia's internals for either:

  • Reproduce it yourself, no secret needed. Every pack lists its artifacts and a merkle root computed from them. Anyone you share the pack with can recompute that root from the listed artifacts and confirm it matches. If a single artifact was changed, added, or removed, the recomputed root will not match, so tampering is detectable without any Dralvia key. The verify response reports this as pack_integrity / merkle_consistent, and the export manifest carries the same kind of anchor (artifacts_merkle_root).
  • Confirm Dralvia signed it. The pack also carries a Dralvia signature. That check confirms Dralvia produced the pack; it is separate from the reproducible check above.

For sharing, the verify response includes a compact trust_receipt: a small artifact with the surface, merkle root, and verification outcome, and no secret, that you can hand to a customer or auditor so they can confirm the evidence themselves. Every module (URL and phishing, smart contract, repository, email, web access, and agent surfaces) produces packs with the same anchors, so the same steps work no matter which scan created the pack.

Is every module equally mature?​

No. Some surfaces are stronger than others.

  • solid: URL & phishing scanner, evidence, reports
  • pilot: pre-sign, repo scanner, email, SWG, identity
  • experimental: long-tail pilot modules that are still being validated

Can this help with AI-assisted attacks?​

Yes, especially where AI speeds up deception, approvals, or workflow misuse. The point is to improve the trust decision before a human or AI agent acts.

Can we use this with AI agents?​

Yes, but the current stance is guardrail-first. The platform is useful before an agent opens a page, calls a tool, signs a transaction, or follows an untrusted path.

What should we validate first as a design partner?​

Do not validate every module at once.

Start with:

  1. suspicious link triage
  2. wallet and pre-sign review
  3. repo and CI gate
  4. evidence/report export

What is the right proof of value?​

One painful real workflow, one clear decision improvement, and one artifact another human can trust.

Day-2 operations​

  • Reuse EvidencePacks and reports in normal handoff flows.
  • Compare whether the team now moves faster or with better confidence.
  • Review where the platform still needs human confirmation.

Self-check playbook​

  • Did the result explain why it mattered?
  • Could another team understand the output without asking the original operator?
  • Did the workflow create a usable artifact for later review?
  • Did the platform reduce uncertainty or just add another score?

What each button does​

  • Export creates an artifact for another team, process, or audit trail.
  • Copy Shareable Link preserves the result in a reusable form.
  • Send to Security moves the result into the security review path.
  • Open Platform continues the workflow inside the main workspace.

Troubleshooting​

  • If the output feels too technical, compare the explanation panels and persona sections before judging the raw payload.
  • If a result looks incomplete, confirm whether the platform is showing a confidence boundary rather than failing silently.
  • If you need a stronger artifact, use the report or EvidencePack flow instead of a screenshot.

API and automation​

  • Use the public scan and contract endpoints for repeatable ingestion.
  • Use exports when another system expects JSON, SARIF, SBOM, or report bundles.
  • Start with one automation path, not every possible integration at once.

API error quick reference​

  • 401 Unauthorized: the session or API credential used for a scan or export is missing or invalid.
  • 403 Forbidden: the current role is not allowed to access the requested workflow, artifact, or export type.
  • 404 Not Found: the referenced result, report, or EvidencePack does not exist in the current workspace scope.
  • 429 Too Many Requests: the workspace exceeded a workflow or export rate limit; wait before retrying.
  • 500 Internal Server Error: the backend failed while building or exporting the requested evidence; retry once and capture the timestamp if it happens again.

Next best actions​

Related guides:

  • EvidencePack verification

  • URL & phishing scanner

  • Pricing & Plans

  • If explanation quality is strong enough, test the same workflow with a second real input.

  • If the evidence model is strong enough, validate the report or EvidencePack handoff with another team.

  • If a module is still early, keep it in pilot scope instead of forcing broader rollout.

FAQ​

Is Dralvia trying to replace every security product?​

No. The strongest current value is around evidence-backed trust decisions across risky digital actions.

Is every score final?​

No. Some surfaces intentionally expose confidence and deeper-review boundaries.

What matters more: score or explanation?​

Explanation, evidence, and the next action are usually more important than the number alone.

Next steps​

  • Run one real workflow.
  • Export one artifact.
  • Decide whether the product is reducing uncertainty for your team.

Known limits and rate limits​

  • Plan quotas apply: scan, render, report, repository, and API-key limits follow your plan tier. See Pricing & Plans for the enforced defaults.
  • API requests are rate limited per workspace. A 429 Too Many Requests response means a rate or quota window was exceeded; wait for the reset window, then retry once and reduce burst volume.
  • Free Evaluation has the tightest limits and is subject to anti-abuse controls. Paid plans raise the limits; Enterprise sets custom limits by contract.