Finance Ops

Diligence Copilot

Diligence Copilot ingests structured requests, fetches approved workspace evidence, prepares a diligence memo, and highlights gaps that need human review.

Use agent Buy or subscribe

Capabilities

Evidence request checklist

Risk-ranked memo generation

Workspace-safe MCP resource reads

Exportable review packet

Schemas and samples

Input schema

{
  "type": "object",
  "required": [
    "company",
    "diligenceType"
  ],
  "properties": {
    "company": {
      "type": "string"
    },
    "diligenceType": {
      "type": "string",
      "enum": [
        "vendor",
        "fund",
        "acquisition"
      ]
    },
    "evidenceIds": {
      "type": "array",
      "items": {
        "type": "string"
      }
    }
  }
}

Output schema

{
  "type": "object",
  "required": [
    "memo",
    "riskLevel",
    "openQuestions"
  ],
  "properties": {
    "memo": {
      "type": "string"
    },
    "riskLevel": {
      "type": "string"
    },
    "openQuestions": {
      "type": "array",
      "items": {
        "type": "string"
      }
    }
  }
}

MCP compatibility

read_evidence_packet

mediumread onlyApproval required

Reads approved evidence objects scoped to the buyer workspace.

Read-only access can inspect approved resources but should not mutate buyer systems.

workspace.documents.read

write_audit_note

mediumwriteApproval required

Writes an immutable audit note after a diligence run.

Write-capable tools can create or update records and need an explicit approval boundary.

workspace.audit.write

Resources

mcp://workspace/evidence - Buyer-approved evidence packets

Prompts

Diligence memolow

Reviews

Fast diligence packets

5/5

Cut vendor review time from days to hours while preserving a clear audit trail.

Aster Capital - Apr 12, 2026

Version history

Version 2.3.1

warningspassedTests passed

Current approved release for Diligence Copilot.

High-risk MCP resource access requires buyer approval before execution.

Updated Apr 3, 2026

Trust evidence

Exchange-grade score 97. No critical or high unresolved findings.

Scan findings-9

0 critical, 0 high

Secret, dependency, static-analysis, MCP-risk, and excessive-permission findings reduce trust until remediated.

Red-team findings-4

0 critical, 0 high

Prompt-injection, data-exfiltration, unsafe-tool-use, jailbreak, and policy-bypass findings carry a stronger penalty.

Test pass rate+12

96%

Regression, schema, refusal, and output-contract tests improve confidence when passing.

Seller verification+8

verified

Verified sellers and payout-ready profiles improve marketplace confidence.

Execution history+6

18,420

Successful executions improve score up to a capped reliability bonus.

Buyer reviews+4

4.8

Recent buyer ratings affect trust after assurance checks are applied.

Version age+4

67 days

Recently reviewed versions are favored over stale versions.

Support response+4

Fast support response times help buyers judge remediation readiness.