AGENT SKILL

signet-guardian

Payment guard middleware for AI agents. Use whenever any skill is about to initiate a payment. Runs a preflight check against the user's policy (payments enabled, per-transaction limit, monthly cap). Returns ALLOW, DENY, or CONFIRM_REQUIRED. Other payment-capable skills must call signet-preflight before proceeding and signet-record after a successful payment.

License 许可
MIT
最近更新
2026-06-07
Stars
★ 0
Forks
1

01 前言 Front Matter

name
signet-guardian
metadata
{ "openclaw": { "emoji": "🛡️", "homepage": "https://getsignet.xyz" } }
description
Payment guard middleware for AI agents. Use whenever any skill is about to initiate a payment. Runs a preflight check against the user's policy (payments enabled, per-transaction limit, monthly cap). Returns ALLOW, DENY, or CONFIRM_REQUIRED. Other payment-capable skills must call signet-preflight before proceeding and signet-record after a successful payment.

03 SKILL.md 正文

Signet Guardian — Payment Guard Middleware

Overview

Signet Guardian is a policy firewall for money actions. It does not intercept payments at runtime by itself; payment-capable skills must route through it by contract:

  1. Before any payment: call signet-preflight (amount, currency, payee, purpose).
  2. If result is ALLOW or CONFIRM_REQUIRED (and user has confirmed): the skill may proceed.
  3. If result is DENY: do not proceed; tell the user the reason.
  4. After a successful payment: call signet-record to append to the ledger.

This gives one place to enforce: master switch (payments on/off), max per transaction (e.g. £20), max per month (e.g. £500), and optional confirmation above a threshold (e.g. £5).

Concurrency: Preflight is advisory (no lock). Record enforces the monthly cap under a file lock ({baseDir}/references/.ledger.lock): it re-checks the cap before appending and refuses to record if the month would be exceeded. So the monthly limit is enforced at record time; idempotency and cap are both safe under concurrent calls. Preflight can still be used to fail fast; the definitive check is in record.

Currency: No FX conversion. The request currency must match the policy currency; otherwise preflight returns DENY. Conversion source/rules are not defined.

Policy (user configuration)

Source of truth: OpenClaw config first (signet.policy in the main config, e.g. editable in the Control UI if the extension is installed), then fallback to {baseDir}/references/policy.json. OpenClaw sets {baseDir} via OPENCLAW_SKILL_DIR or OPENCLAW_BASE_DIR.

Field Meaning
paymentsEnabled Master switch. If false, all payments are denied.
maxPerTransaction Max amount allowed for a single transaction (e.g. 20).
maxPerMonth Max total spend in the current calendar month (e.g. 500).
currency ISO currency code (e.g. GBP, USD). Request currency must match.
requireConfirmationAbove Above this amount, return CONFIRM_REQUIRED so the user must explicitly confirm (e.g. 5).
blockedMerchants Optional list of substrings; payee matching any is denied.
allowedMerchants Optional; if non-empty, only payees matching one of these are allowed.
version Optional number for future policy migrations.

Default behaviour: If the policy file is missing or invalid, preflight returns DENY (default-deny).

Commands

signet-preflight

Run before initiating any payment. Validates: payments enabled, currency match, amount > 0 and ≤ max per transaction, (current month spend + amount) ≤ max per month, and optional merchant rules. Optionally requires explicit confirmation above a threshold. Amount must be greater than zero.

signet-preflight --amount 15 --currency GBP --payee "shop.example.com" --purpose "Subscription"

Optional:

  • --idempotency-key "unique-key" — Used when recording later to avoid duplicate ledger entries.
  • --caller-skill "skill-name" — Name of the skill invoking the guard (for audit).

Output (JSON):

  • { "result": "ALLOW", "reason": "Within policy" } — Proceed with the payment.
  • { "result": "CONFIRM_REQUIRED", "reason": "..." } — Ask the user for explicit confirmation; if they agree, proceed then call signet-record. (Confirmation is the caller’s responsibility.)
  • { "result": "DENY", "reason": "..." } — Do not proceed. Notify the user.

Every DENY is logged to the audit trail.

Exit code: 0 for ALLOW or CONFIRM_REQUIRED, 1 for DENY.

signet-record

Call after a payment has successfully been made. Appends one line to the ledger (append-only). If an idempotency key was used in preflight, pass the same key here to avoid double-counting.

Record validation scope: signet-record re-checks only currency and monthly cap (under lock). It does not re-check paymentsEnabled or merchant allow/block lists. Policy enforcement (switch, merchants, per-tx limit) is done at preflight (and in an optional future authorize phase). Record is the post-success log; the cap check at record time prevents double-counting when concurrent preflights both allowed.

signet-record --amount 15 --currency GBP --payee "shop.example.com" --purpose "Subscription" --idempotency-key "sub-123"

Optional: --caller-skill "skill-name" for audit.

If the same idempotency-key was already recorded, the command is a no-op (idempotent).

signet-report

Shows spending and transaction history for the user.

signet-report --period today
signet-report --period month

signet-policy

Show, edit, or configure policy via wizard.

signet-policy --show    # Print current policy (config, then file)
signet-policy --edit    # Open policy.json in $EDITOR
signet-policy --wizard  # Interactive step-by-step setup (no JSON)
signet-policy --migrate-file-to-config  # One-time: copy file policy into OpenClaw config

Audit (ledger and deny log)

Ledger file: {baseDir}/references/ledger.jsonl. Format is strict JSONL: one JSON object per line, newline-separated (no space between entries). Each line contains:

  • ts — Timestamp UTC (ISO 8601).
  • callerSkill — Optional; skill that invoked preflight/record.
  • idempotencyKey — Optional; dedupe key for record.
  • status — completed or denied.
  • reason — Decision reason (especially for denials).
  • Plus: amount, currency, payee, purpose.

All preflight denials are appended to the same ledger with status: "denied" and a reason.

Critical Rules (for the agent)

  1. Never skip preflight — Any payment from any skill must go through signet-preflight first. No exceptions.
  2. Respect DENY — If preflight returns DENY, do not attempt the payment. Tell the user the reason.
  3. CONFIRM_REQUIRED — If preflight returns CONFIRM_REQUIRED, ask the user explicitly (“Allow this payment of £X to Y?”). Only proceed if they confirm, then call signet-record.
  4. Always record success — After a successful payment, call signet-record with the same amount, currency, payee, purpose, and idempotency key (if used).
  5. Idempotency — For critical flows, use a stable --idempotency-key (e.g. order ID or request ID) so retries do not double-count in the monthly total.
  6. Default-deny — If the policy file is missing or corrupt, the skill denies by default.
  7. Record is authoritative for cap only — The monthly cap is enforced when recording (under lock). If signet-record fails with a cap error, the payment already happened; do not retry without user confirmation. For cap-safe flows before payment, a future authorize (reservation under lock) then settle (convert reservation to completed) pattern can reserve budget before the payment is made.

First Run

On first use, the user must have a valid {baseDir}/references/policy.json. Run signet-policy --show to see current policy; if missing, create it (e.g. via signet-policy --edit) with at least:

  • paymentsEnabled: true/false
  • maxPerTransaction: number
  • maxPerMonth: number
  • currency: e.g. "GBP"
  • requireConfirmationAbove: number (e.g. 5)

Ledger lives at {baseDir}/references/ledger.jsonl; no extra setup required.

04 同仓库更多 Skills

javimosch
javimosch/open-claw-skills

crowd-prompting

A marketplace where AI agents improve prompts, system instructions, tool descriptions, and other text-based content with domain expertise from real-world operations — and earn tokens for valuable contributions.

★ 0
javimosch
javimosch/open-claw-skills

ceaser

Interact with the Ceaser privacy protocol on Base L2 using the ceaser-mcp MCP tools. This skill uses the ceaser-mcp npm package for ALL operations -- shield, unshield, note management, and protocol queries. All ceaser tool calls use CLI subcommands (npx -y ceaser-mcp <subcommand>). Ceaser is a privacy-preserving ETH wrapper using Noir/UltraHonk zero-knowledge proofs on Base L2 (chain ID 8453).

★ 0
javimosch
javimosch/open-claw-skills

ceaser-send

Fully automated private ETH transfer via Ceaser Protocol on Base L2 using the ceaser-mcp MCP tools. This skill uses the ceaser-mcp npm package for ALL operations -- all ceaser tool calls use CLI subcommands (npx -y ceaser-mcp <subcommand>). WARNING -- creates observable on-chain links between user wallet and hot wallet (see Privacy Warning). Generates an ephemeral hot wallet with BIP-39 mnemonic recovery, user funds it, agent signs and broadcasts the Shield TX automatically. Exactly one manual step required (funding the hot wallet with ETH). Uses Noir/UltraHonk zero-knowledge proofs.

★ 0
javimosch
javimosch/open-claw-skills

dilemm-ai

Compete in DilemmAI, the prisoner's dilemma AI arena at dilemm.ai. Use when an OpenClaw agent wants to create an account, design and submit strategy prompts for their bot, enter matchmaking, analyze historical matches and leaderboards, and iterate on bot strategy. Requires headless browser automation (no API available yet).

★ 0
javimosch
javimosch/open-claw-skills

moldium

Post and manage content on the Moldium blog platform. Triggered by "post to Moldium", "write a blog post", "publish an article", etc.

★ 0
javimosch
javimosch/open-claw-skills

arxiv-paper-reviews

Interact with arXiv Crawler API to fetch papers, read reviews, submit comments, search papers, and import papers. Use when working with arXiv papers, fetching paper lists by date/category/interest, viewing paper details with comments, submitting paper reviews, searching papers by title, or importing papers from arXiv URLs via API at http://weakaccept.top:8000/.

★ 0