界面已本地化;技能介绍与说明文档按所选语言由 AI 自动翻译,代码、命令、路径和专业术语保留原文,以原作者文档为准。
探索技能
开发与工程 · SKILL.md

settlement-witness

settlement-witness

【开发与工程】OpenClaw skill: fetch replay-stable PASS/FAIL receipts from the public SettlementWitness HTTPS endpoint. Stateless. Deterministic. Signed.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准

技能说明

GitHub

name: settlement-witness description: "OpenClaw skill: fetch replay-stable PASS/FAIL receipts from the public SettlementWitness HTTPS endpoint. Stateless. Deterministic. Signed."

SettlementWitness (OpenClaw Skill)

This skill calls the public SettlementWitness HTTPS endpoint. No local services are required.

What this skill does

When a task is complete and you need proof-of-delivery before settlement, call SettlementWitness to obtain a replay-stable receipt.

SettlementWitness does not judge quality or arbitrate disputes. It only checks whether the provided output matches the provided spec under deterministic rules.

Canonical endpoints

Required input shape

Provide:

  • task_id (string)
  • spec (object)
  • output (object)

Example request

{ "task_id": "example-002", "spec": { "expected": "foo" }, "output": { "expected": "foo" } }

How to interpret

  • If verifier_response.verdict == PASS: verified completion
  • If verifier_response.verdict == FAIL: not verified (do not settle automatically)
  • receipt_id is the stable identifier to store/log/share

Safety notes

  • Never send secrets (private keys, API keys) in spec/output.
  • Keep spec/output minimal and deterministic (hashes/IDs are ideal).

Install (for OpenClaw users)

Copy this folder into your OpenClaw skills directory as: settlement-witness/SKILL.md

我的收藏