← Back to Skills Marketplace
ivangdavila

Banking

by Iván · GitHub ↗ · v1.0.0
darwinlinuxwin32 ✓ Security Clean
477
Downloads
0
Stars
4
Active Installs
1
Versions
Install in OpenClaw
/install banking
Description
Manage retail and business banking workflows with payment operations, account controls, reconciliation, fraud response, and compliant communication.
README (SKILL.md)

Setup

On first use, read setup.md for activation boundaries and context capture priorities.

When to Use

Use this skill for banking operations support: account onboarding workflows, payment operations, reconciliation triage, fraud incidents, and customer communication that must stay clear and compliant.

Architecture

Memory lives in ~/banking/. See memory-template.md for structure and status fields.

~/banking/
|-- memory.md                 # Status, activation scope, operating context
|-- incidents.md              # Open fraud and operations incidents
|-- payment-controls.md       # Verified controls by rail and account type
`-- communication-notes.md    # Approved customer messaging patterns

Quick Reference

Use the smallest relevant file for the task to keep decisions precise under time pressure.

Topic File
Setup process setup.md
Memory template memory-template.md
Request intake and classification intake-checklist.md
Payment rails and controls payment-ops.md
Fraud and outage handling incident-response.md
Customer-safe wording customer-messaging.md
Regulatory and legal boundaries compliance-scope.md

Core Rules

1. Classify the Request Before Giving Steps

  • Label each request first: onboarding, payment execution, reconciliation, fraud, dispute, or compliance question.
  • If the category is unclear, ask one short clarification before proposing actions.

2. Confirm Jurisdiction and Account Context

  • Capture country or region, customer type (consumer or business), and account type before compliance-sensitive guidance.
  • Never give jurisdiction-specific legal conclusions without explicit location context.

3. Use Control-First Payment Guidance

  • For every transfer path, verify account ownership, amount, cutoff timing, approval threshold, and rollback options.
  • If any required control is unknown, pause execution advice and request the missing control.

4. Treat Incidents as Containment Then Recovery

  • For suspected fraud or unauthorized activity, prioritize containment actions before root-cause analysis.
  • Keep incident actions timestamped and reversible where possible.

5. Keep Communication Clear, Neutral, and Accurate

  • Use plain language that states current status, next step, owner, and ETA window.
  • Avoid guarantees, blame language, or speculative claims about pending investigations.

6. Keep Memory Actionable and Verifiable

  • Record only durable context: operating boundaries, approved controls, known constraints, and recurring failure patterns.
  • Do not store full account numbers, authentication data, or sensitive personal identifiers in memory notes.

7. Escalate High-Risk or Restricted Requests

  • Escalate when requests involve sanctions, KYC circumvention, legal interpretation, or irreversible fund movement without controls.
  • Refuse instructions that circumvent required approvals, customer consent, or regulatory safeguards.

Common Traps

  • Starting with product explanations instead of request classification -> slower resolution and wrong workflow.
  • Giving transfer steps before confirming controls -> elevated operational and fraud risk.
  • Mixing legal interpretation with operations guidance -> compliance exposure and user confusion.
  • Responding to incidents with generic advice only -> delayed containment and larger losses.
  • Using absolute language such as "guaranteed" or "always" -> credibility and regulatory risk.
  • Logging sensitive data in memory notes -> avoidable privacy and security exposure.

Data Storage

  • Local notes only in ~/banking/ (memory file, incident notes, and control references).
  • Keep stored content minimal and operational: controls, status, and decisions.
  • Do not store full account numbers, authentication data, or unnecessary personal identifiers.

Security & Privacy

Data that leaves your machine:

  • None by default. This skill is instruction and workflow guidance only.

Data that stays local:

  • Operational context and notes in ~/banking/.

This skill does NOT:

  • Access bank portals or execute fund transfers automatically.
  • Request undeclared network calls.
  • Store authentication data or full account numbers in memory files.
  • Modify files outside ~/banking/ for storage.
  • NEVER modifies its own skill definition file.

Related Skills

Install with clawhub install \x3Cslug> if user confirms:

  • payments - Payment workflows and transaction operations patterns.
  • accounting - Ledger, reconciliation, and financial reporting support.
  • invoice - Invoice lifecycle workflows and settlement tracking.
  • money - Personal money management and budgeting fundamentals.
  • invest - Investment analysis workflows for portfolio decisions.

Feedback

  • If useful: clawhub star banking
  • Stay updated: clawhub sync
Usage Guidance
This skill is internally coherent and appears to be a local-playbook for banking operations. Before installing or using it with real cases: (1) Restrict the filesystem permissions on ~/banking/ so only authorized agents/users can read it; (2) enforce your own rules to never store full account numbers, credentials, or other PII in that folder; (3) confirm that other installed skills or the agent runtime do not have network access or credentials that could cause data in ~/banking/ to be exfiltrated; (4) consider testing with dummy data first; and (5) if you expect the agent to perform real transfers or access bank portals, choose a different skill that explicitly requests the necessary credentials and integrations.
Capability Analysis
Type: OpenClaw Skill Name: banking Version: 1.0.0 The OpenClaw Banking skill bundle is benign. All files, including the SKILL.md and other markdown instructions, consistently emphasize security, compliance, and conservative operational guidance. SKILL.md explicitly states that the skill does not make network calls, access bank portals, execute fund transfers automatically, or modify files outside `~/banking/`. Furthermore, it instructs the AI agent to refuse requests that circumvent approvals, customer consent, or regulatory safeguards, and to escalate high-risk situations. There is no evidence of prompt injection, data exfiltration, malicious execution, or persistence mechanisms.
Capability Assessment
Purpose & Capability
Name/description (banking operations, incident response, communications) match what the skill requests and instructs: a local memory/config path (~/banking/) and structured guidance files. No unrelated binaries, cloud credentials, or network integrations are requested.
Instruction Scope
SKILL.md and the included docs confine activity to guidance and local notes in ~/banking/ (memory, incidents, controls, messaging). Instructions explicitly forbid storing full account numbers/authentication and disallow network calls. This is coherent; note that enforcement depends on the agent/runtime following the guidance.
Install Mechanism
No install spec and no code files — instruction-only — so nothing is downloaded or written to disk by an installer. Lowest-risk install profile.
Credentials
The skill declares only a config path (~/banking/) and no environment variables or credentials. That is proportionate for a workflow/reference skill that does not itself access bank APIs or execute transfers.
Persistence & Privilege
The skill persists local state in ~/banking/, which is reasonable for incident history and templates. It is not always-enabled and does not request elevated system-wide privileges. Caveat: because it writes persistent notes about banking incidents, the user should control filesystem permissions and be careful what sensitive data is saved — the skill's guidance attempts to limit sensitive storage but cannot enforce it itself.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install banking
  3. After installation, invoke the skill by name or use /banking
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release with structured banking workflows for intake, operations, incident response, and customer-safe messaging.
Metadata
Slug banking
Version 1.0.0
License
All-time Installs 4
Active Installs 4
Total Versions 1
Frequently Asked Questions

What is Banking?

Manage retail and business banking workflows with payment operations, account controls, reconciliation, fraud response, and compliant communication. It is an AI Agent Skill for Claude Code / OpenClaw, with 477 downloads so far.

How do I install Banking?

Run "/install banking" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Banking free?

Yes, Banking is completely free (open-source). You can download, install and use it at no cost.

Which platforms does Banking support?

Banking is cross-platform and runs anywhere OpenClaw / Claude Code is available (darwin, linux, win32).

Who created Banking?

It is built and maintained by Iván (@ivangdavila); the current version is v1.0.0.

💬 Comments