SkyInsights Risk Intelligence

Use {skillDir}/scripts/skyinsights.py for all API calls. The script handles credentials, request formatting, error handling, screening polling, and terminal-friendly output.

Running Commands

python {skillDir}/scripts/skyinsights.py <subcommand> [args...]

Subcommands

• kya <address> [chain=eth] — GET /v4/kya/risk
• labels <address> [chain=eth] — GET /v4/kya/labels
• screen <address> [chain=eth] [rule_set_id=standard-mode-rule-set] — GET /v4/kya/screening_v2 + poll (⚠️ 5–15s)
• kyt <txn_hash> <chain> — GET /v4/kyt/risk
• help — show usage

rule_set_id options: standard-mode-rule-set (default), fast-mode-rule-set

Supported Chains

Per-endpoint chain support. Use the API Value when specifying the chain parameter.

kya/screening_v2 supports: btc eth polygon op arb avax bsc tron wemix base multi-chain is supported by kya/risk only.

Default Workflow

Route user requests as follows:

• User mentions a transaction hash / tx / txn → run kyt
• User asks about labels / entity / institution / exchange for an address → run labels
• User asks if an address is risky / safe / suspicious (no subcommand specified):
  1. Run kya first — fast, checks the address's own labels and risk score.
  2. If kya returns None or Low, automatically follow with screen — a clean address can still have counterparty exposure to sanctioned or hacked funds. Tell the user: "该地址自身风险较低，正在进行合规筛查以检查对手方暴露情况，请稍候（约 5–15 秒）…"
  3. If kya returns Medium, High, or Unknown, present the result directly without running screen.

Output Expectations

Printed shapes per subcommand:

# kya
Risk: {None|Low|Medium|High|Unknown} (score={0-5})
Reasons: {reason1}, {reason2}     ← omitted if empty

# labels
Entity: {name}  Type: {type}
Labels: {label1}, {label2}        ← omitted if none

# screen
Screening: {Pass|Fail|Pending}
Flagged counterparties: {count}
{counterparty_address}  {risk_level}  {reason}  ← one line per flagged party