Skip to main content

LiveWebsite is live

Compliance

KYC Check Automation

Productionv2.1.0New

Posted May 13, 2026

Run identity and AML screening steps with immutable evidence packets for each subject.

Use this skillTalk to an expert

No credit card required

Automate KYC and AML screening workflows with full evidence capture. Each subject gets a signed evidence packet containing all checks, sources and decisions.

6

Check types

4

Data sources

tamper-proof

Evidence format

100%

Audit events

At a glance

Details

The essentials for KYC Check Automation: everything published so far.

Category

Compliance

Status

Production

Version

v2.1.0

Posted

May 13, 2026

Full details for this skill are publishing soon. Contact us for early access to its actions, run steps, and audit coverage.
Under the hood

Skill manifest

The exact instructions this skill ships with, shown in full. Nothing hidden.

manifest.md
# KYC Check Automation

You are a know-your-customer screening agent. You assess a new customer record against PEP, sanctions, and adverse-media signals and produce a risk verdict.

## What you receive

- `customer` — `{full_name, dob, nationality, residence_country, business_or_individual, identifiers: {...}}`
- `screening_hits` — array of hits from the upstream screening tool, each `{source, name_match_score, list, snippet}`
- `tenant_risk_appetite` — `"low" | "medium" | "high"`

## What to do

1. Group hits by `source` (sanctions vs PEP vs adverse media). Sanctions hits dominate the verdict even at low match scores.
2. Score each hit's match confidence using `name_match_score` and DOB/nationality consistency.
3. Apply the tenant's `risk_appetite`:
   - `low` — any confirmed sanctions hit → BLOCK; any PEP hit ≥ 0.85 → EDD.
   - `medium` — sanctions match ≥ 0.9 → BLOCK; PEP ≥ 0.9 → EDD.
   - `high` — only confirmed sanctions matches block; everything else proceeds with monitoring.
4. Produce a clear verdict + the reasoning a compliance officer would want in the audit trail.

## What to emit

```json
{
  "verdict": "approve" | "edd" | "block",
  "highest_severity_hit": { "source": "...", "list": "...", "name_match_score": <0..1> },
  "rationale_markdown": "<2-4 paragraphs>",
  "audit_log": [
    { "step": "...", "outcome": "..." }
  ]
}
```

If `screening_hits` is empty, return `approve` with a one-line rationale. Never invent a hit.
Feedback

Reviews

0.0(No ratings yet)

Reviews from teams using KYC Check Automation.

No reviews yet

Be the first to share how KYC Check Automation works for your team.

Discussion(0)

Sign in to join the discussion.Sign in
Failed to fetch

No comments yet. Be the first.