SwiftAdviser

Mandate Skill — AI skill for Claude Code

AI community

Wallet-agnostic non-custodial policy guardrails for AI agents.

How to install Mandate Skill

This entry records only its repository, not the path inside it, so there is no exact command to give. Open SwiftAdviser/mandate-skill and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Mandate Skill does

Wallet-agnostic non-custodial policy guardrails for AI agents.

Alternatives in AI

  • Opengap — A framework-agnostic, git-native standard for defining AI agents 2.9k ★
  • Open Science — Open Science Desktop — local-first, model-agnostic AI research workbench for macOS, Windows & Linux 1.5k ★
  • Paritok 4b V1 — Non-destructive compression gateway for AI coding agents 1.4k ★

README

Mandate.md

demo website skills.sh license

Mandate Skill

**Approve intent, not just transactions.**

*Transaction intelligence and control for autonomous agents, delivered as an agent skill.*

[One-minute demo](https://www.youtube.com/watch?v=EwZdtncc5wQ)

This repository packages Mandate as a portable **agent skill**. Drop it into any skill-aware coding agent (Claude Code, Codex, Cursor, OpenCode, Gemini CLI, Windsurf, and 35+ others) and your agent gains a reason-aware control layer that evaluates **why** a payment is being made before it signs. For hook-based enforcement, pair this skill with the OpenClaw or Claude Code plugin (see Install below).

Why it matters

  1. Intent-aware payment decisions. Evaluate why an agent wants to pay, then approve, block, or escalate before signing.
  2. Real-time risk prevention. Stop fraud, prompt-injection payments, and costly mistakes before funds move.
  3. Complete payment auditability. Full audit trail of every decision with amount, timing, and rationale.

Install

Universal install via [skills.sh](https://skills.sh) (40+ agents, no hooks)

Installs `SKILL.md` into Claude Code, Codex, Cursor, OpenCode, Gemini CLI, Windsurf, and 35+ other agents:

npx skills add SwiftAdviser/mandate-skill

Flags: `-g` for global (`~/.claude/skills/...`), `-a claude-code` to target a specific agent, `-y` to skip prompts.

No hooks layer, so the agent must volun

Related Skills

Agent Context — AI skill for Claude Code

Agent Context

Layered, agent-agnostic context architecture for AI coding agents: a tiny always-on baseline, on-demand skills

AI community
Stack Ia Dev — AI skill for Claude Code

Stack Ia Dev

Tool-agnostic AI software engineering stack: deterministic safety guardrails, project-context profiling, multi

AI community
Agentic Shopping Site — AI skill for Claude Code

Agentic Shopping Site

Landing site for Agentic Shopping Autopilot — open-source AI shopping agent that reads your notes first, then

AI community
Venice Py — AI skill for Claude Code

Venice Py

Unofficial async-first Python SDK for the Venice.ai API — chat, image, video, music, audio, embeddings, TEE at

AI community
Emblem AI Agent Wallet — AI skill for Claude Code

Emblem AI Agent Wallet

Multi-chain crypto wallet management across 7 blockchains (Solana, Ethereum, Base, BSC, Polygon, Hedera, Bitco

AI community
Soulpass CLI Skill — AI skill for Claude Code

Soulpass CLI Skill

Hardware-secured Solana wallet, trading terminal, and agent identity layer for AI agents. Jupiter DEX swaps, m

AI community

Related Agents

Attack Kitty — Development agent for Claude Code

Attack Kitty

Non-author verification and review expert — fresh-context judgment under a thin mandate. Twelve mandate types

Development community
Reagent Orchestrator — Project Management agent for Claude Code

Reagent Orchestrator

Project process orchestrator — enforces policy.yaml autonomy level, routes tasks to specialist agents, checks

Project Management community
Rea Orchestrator — Development agent for Claude Code

Rea Orchestrator

Primary routing agent — enforces .rea/policy.yaml autonomy level, checks HALT before delegation, and routes no

Development community