From payments-fintech-compliance
Drafts the incident-review pack for a payments-operations event at a fintech program operator, BaaS platform, money transmitter, neobank, wallet, BIN sponsor, or payments processor. The pack carries an incident summary, customer-impact population, named-rail and sponsor-bank notification triggers, root cause, affected transaction population by rail, remediation actions, sponsor-bank reporting, and the regulator-facing artifact list. Output is review-ready for the program operator's second line and for production to the sponsor bank, the named payment-rail authorities (where applicable), and the regulator-facing incident file. Best for: - An ACH return spike, mis-posted batch, double-debit, stuck FedNow / RTP transfer, faulty Reg E claim queue, card-network fraud-rate or chargeback program escalation, or processor-side outage has happened and second line owns the review pack. - A sponsor-bank annual review, internal audit, or examiner data request includes incident retrospectives and the team needs the structured artifact. - A subcontractor (processor, ledger provider, KYC vendor, fraud-decision vendor) outage has cascaded into the program operator's customer-impact surface and the team needs the rail-by-rail attribution and notification map. - A near-miss has been escalated and second line wants the same structure to evidence the control catch even though no customer was impacted. Not the right tool when: - The work is a top-down rail-and-segment risk assessment (use `payments-risk-assessment`). - The work is a controls inventory or sponsor-bank-facing self-evidence pack (use `fintech-partner-controls`, or `open-banking-data-controls` where the surface is §1033 data flows). - The incident is purely cyber with no payments-operational impact and the disclosure-track artifact is what is needed (use `risk-reporting/cyber-disclosure-readiness`). This skill defers the public-disclosure leg to that one and stays focused on the payments-operational chain. - The work is a SAR-decision review (use `financial-crime-governance/sar-decision-qa`). - The work is enterprise-wide incident management not anchored to payments rails (use the generic incident pattern in `risk-compliance-core` / `compliance-testing`).
How this skill is triggered — by the user, by Claude, or both
Slash command
/payments-fintech-compliance:payment-operations-incident-review [incident reference, rails affected, customer-impact magnitude, source posture, audience][incident reference, rails affected, customer-impact magnitude, source posture, audience]The summary Claude sees in its skill listing — used to decide when to auto-load this skill
The pack is what a program-operator second line, a sponsor-bank operating committee, an internal auditor, or an examiner reaches for when a payments incident lands. The artifact has named sections — incident summary, chronology, customer impact, affected transaction population, regulatory notification triggers, root cause, remediation, sponsor-bank reporting, network reporting, regulator-facing...
TROUBLESHOOTING.mdexamples/fednow-outage-payroll-on-demand-fallback-to-same-day-ach.mdexamples/mis-applied-ach-return-batch-reg-e-timing-exposure.mdreferences/cross-cutting/cyber.mdreferences/cross-cutting/privacy.mdreferences/sector-overlays/payments-fintech.mdreferences/source-anchors.mdtemplates/default-output.mdThe pack is what a program-operator second line, a sponsor-bank operating committee, an internal auditor, or an examiner reaches for when a payments incident lands. The artifact has named sections — incident summary, chronology, customer impact, affected transaction population, regulatory notification triggers, root cause, remediation, sponsor-bank reporting, network reporting, regulator-facing posture, customer-facing posture, lessons learned — and a structured record. Audience is the head of payments compliance, the BSA officer, the head of operations, the program-operator second line, and on the sponsor-bank side the program-management team and the bank's chief compliance officer counterpart.
This is a payments-incident artifact, not a generic operational post-mortem and not a cyber-disclosure pack. The vocabulary is rails, return-reason codes, dispute-timing windows, settlement finality, request-for-return-of-funds, sponsor-bank notification SLA, FBO subledger reconciliation, customer-impact population, provisional credit. Engineering chronology is necessary input but not the centre of gravity; the centre is the rail-by-rail notification map and the sponsor-bank reporting chain. Where the incident has a cyber leg with public-registrant disclosure reach, the pack flags it and cross-references risk-reporting/cyber-disclosure-readiness for the disclosure-track artifact rather than duplicating that work.
The pack is a draft. The named approver — head of payments compliance, BSA officer, sponsor-bank operating committee, head of operations, internal audit on retrospective review — decides.
Most of the spine is set by the incident itself and the source posture. A few things settle before drafting.
risk-reporting/cyber-disclosure-readiness.When scope (per risk-compliance-core/scoping) is supplied, the skill consumes it (institution.type, institution.primary_regulators, sector_overlay_set, cross_cutting_overlay_set, persona.role, source_posture). When it is not, the skill asks the four questions above and defaults to public posture if the practitioner declines. Note in the artifact that scope was not formalised.
The pack has the same spine across incident types. A senior reviewer walks it roughly in the order the incident itself unfolded, not in lockstep. Cite a source from references/source-anchors.md (or a loaded overlay) by file path for every material claim; mark unsupported items [evidence needed] rather than letting them pass.
The incident summary is one paragraph. Rail or rails affected, product, customer-impact magnitude (population count and dollar value, even rough), incident-management status, whether the sponsor-bank notification clock has been triggered, whether any named-rail notification clock is in flight. This is the read securities counsel and the sponsor-bank chief compliance officer counterpart use to triage.
The chronology is timestamped end-to-end from first signal through customer-facing remediation. Each row carries timestamp_local, timestamp_UTC, actor (role, not person; subcontractor name where applicable), action, system or rail touched, evidence reference. The chronology distinguishes what is observed from what is inferred and from what is unknown; the unknowns drive the open-questions block. Subcontractor touchpoints (processor, ledger provider, KYC vendor, fraud-decision vendor, sponsor-bank operations) are explicitly named at each step where they sit in the chain.
The customer impact block names the population count, dollar magnitude, geographic breakdown by US state and corridor where cross-border, and segment breakdown where relevant (consumer, SMB, payroll-on-demand, gig disbursement, BNPL, cross-border remittance, high-risk vertical). It also names how the population was identified (full-population reconciliation against the system-of-record vs sample-based vs complaint-driven) and how the reconciliation closed. A population stated without an identification method is not a finding; it is an assertion.
The affected transaction population is the rail-by-rail breakdown. One row per rail in scope. Columns: volume, value, status (completed / reversed / re-initiated / still in flight / written off), source-system reference. The rail-by-rail view is what the sponsor bank's program-management team reconciles against its own reporting feed and what the named-rail authority sees if a network or scheme reporting line is owed.
The regulatory notification triggers block carries one sub-section per applicable regime. The sub-sections are explicit per source rather than rolled into a single "regulatory" line; the triggers run independently and treating them as one line is the most-frequent finding in this lane.
risk-reporting/cyber-disclosure-readiness for the disclosure-track artifact; the cross-cutting cyber overlay (references/cross-cutting/cyber.md) carries the parallel-clock detail.The root cause block separates trigger from contributing factors and distinguishes control-design failures from control-execution failures. Subcontractor-side root cause is included where the firm has visibility (vendor incident report, vendor-side log access, vendor post-incident review). The block does not assign blame to a vendor without a basis; if visibility is limited, the block records that as a gap and names the vendor-management ask. Configuration questions (the program-agreement clause-set, the in-flight rail set, the geography footprint) sit beside engineering questions (the failed deployment, the misconfigured rule, the missed alert).
The remediation actions block is structured into immediate, short-term, and structural. Each row carries action, owner role, target date, status, and the control reference (the named control in fintech-partner-controls or in payments-risk-assessment that this remediation updates). Control-design changes (a new control needed) are distinguished from control-execution fixes (the existing control did not fire). The block does not approve or commit on behalf of vendors; vendor-side actions are recorded as commitments to be confirmed.
The sponsor-bank reporting block names what was sent, when, by whom, against which program-agreement clause, channel, and current status. A second row records what is still owed (the post-incident review report, the remediation-tracker update, the next operating-committee read-out). The sponsor-bank reporting chain is contractual and runs independently of any regulator-facing chain.
The network or scheme reporting block applies where the rail set requires it: network monitoring program reporting where the incident touched the program threshold, faster-payments scheme operational-incident reporting where the rail's operating procedures require it, NACHA Risk Management Framework reporting where the incident sits inside the framework's scope. Most incidents do not trigger every line; the block records what does and what does not, with the basis.
The regulator-facing posture block lists the regulator-facing artifact set (incident report to the state-MTL via NMLS, response file for a sponsor-bank-mediated examiner inquiry, response file for a CFPB EXAMS inquiry where the larger-participant rule designates the firm). Each artifact carries a named human-review gate before any outbound; the named approver — the head of payments compliance, the BSA officer, securities counsel where the cyber leg has public-disclosure reach, the sponsor-bank-side reviewer — decides.
The customer-facing posture block records the customer-notification language drafted, the channel (in-app, email, statement insert, regulatory-required disclosure form), the population covered, the date, and the status. It also records the error-resolution status (open consumer-EFT claims, provisional-credit posture, fee-reversal log) and any closed-account or hold-release commitments. Customer-facing posture is what a state-AG inquiry or a CFPB Supervisory Highlights desk-review reads; vague language ("we are addressing the issue") is a finding.
The lessons learned and forward controls block ties the incident back to specific controls in fintech-partner-controls or risk cells in payments-risk-assessment. Each row names the control reference, the recommended change (control-design vs control-execution), the owner role, and the target date. The lesson is the artifact's reason to exist; an incident review that does not produce control updates has not closed.
The artifact closes with open items and named follow-ups, gaps, the recommended disposition (closed, monitoring, open-remediation, escalate), and source citations with date. The named-rail rule citations carry the disclaimer "current edition; specific sections to be confirmed against the firm's licensed copy" wherever the named-rail rule is paywalled.
references/source-anchors.md (or a loaded overlay) by file path. Unsupported claims are marked [evidence needed].Pack depth and length scale to the trigger and the audience. A sponsor-bank operating-committee read-out runs longer than a quarterly internal-audit retrospective. A near-miss pack carries the same named sections but with the customer-impact population at zero and the lessons-learned block as the centre of gravity. A multi-rail incident keeps the rail-by-rail breakdown visible at every relevant block; a single-rail incident collapses the unused rail rows to status lines.
Sector overlay loading is fixed (this skill is in the payments-fintech sector flagship; the overlay is references/sector-overlays/payments-fintech.md and is required-on for every engagement). Cross-cutting overlay loading: cyber is required-on for any incident touching processor / ledger / fraud-decision / sanctions-screening / KYC vendor exposure (substantively all of them) and for any incident with a cyber leg; privacy is required-on where the incident touched personal data; conduct is recommended where the customer-impact pattern maps to UDAAP themes (mis-applied holds, account closures, mis-disclosed fees, missed dispute timing).
references/source-anchors.md — citations and excerpts for the named anchors.references/sector-overlays/payments-fintech.md — payments-fintech sector flavour (rail-specific notification mechanics, sponsor-bank notification chain, faster-payments recovery mechanics, named-rail authority reporting lines).references/cross-cutting/cyber.md — cyber-leg parallel clocks, sponsor-bank prudential cyber-notification clock, vendor-stack cyber posture; cross-link to risk-reporting/cyber-disclosure-readiness for the public-disclosure artifact.references/cross-cutting/privacy.md — personal-data leg, state breach-notification clocks, the named GLBA Safeguards incident-response posture, §1033 implications where authorisation flows or developer-interface tokens were involved.references/firm-overlay.md — firm-installed program-agreement clause-set, named runbooks, named approver chain; consumed when present.templates/default-output.md — incident-review pack template (named sections; markdown is the content spec, the deliverable is a Word memo).examples/ — public-source-derived scenarios (mis-applied ACH return batch causing consumer-EFT timing exposure; FedNow outage with payroll-on-demand fall-back to Same Day ACH).TROUBLESHOOTING.md — recurring pitfalls (engineering post-mortem framing; clock-from-detection on the consumer-EFT clock; missed sponsor-bank contractual clock; reusing card-rail recovery template for faster-payments; categorical MSB assertions; closing without forward-control updates).The plugin-level shared references (references/source-map.md, references/policy-control-library.md, references/public-regulatory-scenarios.md) sit at the plugin root and are consulted alongside the skill-level files.
Default to drafting against templates/default-output.md. Render as Word, Excel, PowerPoint, or Markdown when the audience or workflow asks for it; the typical deliverable is a Word memo via the docx skill in the document-skills plugin, with rail-by-rail population tables that can lift to Excel for sponsor-bank reconciliation. The plugin README names document-skills as a companion plugin.
Downstream consumers — payments-risk-assessment (to update the rail / segment / geography baseline after the incident), fintech-partner-controls (to update the program-agreement-mapped controls inventory after the lessons-learned block lands), risk-reporting/cyber-disclosure-readiness (where the cyber leg has public-disclosure reach) — read the named blocks directly. The named-section structure is the cross-skill contract.
npx claudepluginhub anotb/second-line-financial-services --plugin payments-fintech-complianceProvides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Searches MemPalace before answering questions about past work, people, projects, or prior decisions. Returns verbatim stored content instead of guessing from model memory.