Trust, stated plainly.
mcpindex is the layer an agent consults before it acts on a tool. A trust product has to earn the word: by being legible about how it works, what it anchors, and exactly where it stops. This page is that account.
Every verdict comes from a hybrid evaluation: a deterministic conformance probe (does observed behavior match the declared schema) and an adversarial LLM judge (does the description hide instructions, exfiltration, or overclaims). The output is a per-tool decision - ALLOW, DENY, or REVIEW - with dimension verdicts and severity.
The full method, the four-state model, and the graduation gate are documented in /methodology. It is written to be checked, not taken on faith.
Verdict history is hash-chained and timestamped to Bitcoin via OpenTimestamps. Once a block confirms, the trust record for a tool cannot be quietly rewritten. The claim is precise: anchored history exists. It is not a claim about minute-level ordering inside the confirmation window - see the honest limits below.
- · Advisory, not blocking. mcpindex publishes the verdict; your agent or IDE decides whether to act. We do not sit in the call path or proxy your traffic.
- · Public artifacts in. Verdicts are produced from public tool definitions (description + schema) and live probes. The screener evaluates the text you paste - do not paste real secrets into it.
- · No call-time data. We evaluate the tool definition, not your runtime calls. Your arguments and your data never reach us.
- · Open by default. Methodology, quality scoring, and the registry source are public on GitHub. What we assert, you can audit.
Direct and current: mcpindex is pre-SOC 2. We are not going to imply otherwise. The interim posture is the one above - advisory deployment, no call-time data, public method, Bitcoin-anchored history. Formal attestation (SOC 2 Type 2) is on the roadmap when enterprise demand warrants the audit, not before. If you have a specific compliance requirement, tell us what you need and we will answer honestly about where we are.
Data handling is described in our privacy policy.
A trust product earns trust by stating its edges, on every verdict: conformance is monitored, not enforced; confidences are reported but not yet calibrated (calibrated=false at v1); coverage rolls out as the corpus expands (15 of 150 labels to graduation, adversarial cases first). The full contract lives on /methodology - and it changes there first, before the verdict surface does.
Live system status, data freshness, and the incident log are at /status. To report a problem with a verdict or a security concern, email hello@mcpindex.ai.