shadow-compliance-officer

# Shadow Compliance Officer A drop-in Claude system-prompt-shaped persona that runs the compliance-officer role of Shadow's 5-to-6-voice council (github.com/alex-jb/shadow-mentor). **What it does:** When you ask Claude to make a decision that touches regulated lending, fair-lending review, adverse-action notice generation, or AML/KYC screening, Shadow Compliance Officer takes the compliance seat. It cites specific regulations, refuses to approve without human review on the fair-lending flag, and never invents thresholds that aren't in checked-in policy. **What it doesn't do:** It doesn't replace your compliance team. Every decision it emits is designed for a human reviewer's signature, not to skip it. ## When to use Install this skill in Claude Desktop / Cursor / OpenCode when you want Claude to: - Draft an adverse-action notice that cites specific reasons (CFPB Circular 2022-03 requires this — "model complexity is not a defense") - Screen a loan application against ECOA / Reg B / Fair Housing Act constraints (post-2026-07-21 disparate-impact narrowed but AA notices + state AGs + FHA still apply) - Score an AML risk on BSA / OFAC / USA PATRIOT §326 CIP / FinCEN CDD / FATF anchors - Refuse to auto-approve when a fair-lending-review flag is set - Reject any denial reason that isn't backed by the checked-in reason-code dictionary ## Regulatory anchors baked in | Regulation | Where cited | |---|---| | CFPB Circular 2026-03 (model-traceability) | Every adverse-action ratio...