Decision Receipts

Use before or during consequential Agent actions such as deploy, delete, pay, export, access grants, or high-risk decisions when later offline verification of what happened and under which policy matters.

9 stars 3 forksDevOps & Cloud

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add sandbaseai/workbuddy-skill --skill decision-receipts
Manual install
git clone --depth 1 https://github.com/sandbaseai/workbuddy-skill.git /tmp/workbuddy-skill
mkdir -p ~/.claude/skills
cp -r /tmp/workbuddy-skill/skills/decision-receipts ~/.claude/skills/decision-receipts

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Agent Decision Receipts

## What a receipt proves

A log describes an event after the fact and may be edited. A decision receipt is created before or at the point of a consequential action, records the actor, operation, target, governing policy, and input digest, and is signed so tampering breaks verification. It is local evidence that can be checked offline; it is not a hosted notary, a replacement for logs, or a legal opinion.

## Decision 1: receipt or ordinary log?

Mint a receipt only when all three conditions hold:

| Condition | Meaning |
|---|---|
| side-effecting | writes, sends, deploys, deletes, pays, exports, grants access, or changes external state |
| consequential | a wrong call could cost money, violate policy, break a service, or harm a person |
| later-provable | an auditor, reviewer, insurer, regulator, or counterparty may ask what happened and why |

High-signal defaults include deploy, delete, pay/refund, access changes, data egress, claim approval/denial, and high-risk model decisions. Read-only, reversible, trivial actions normally need ordinary trace logs; receipt everything and the signal becomes noise. Record the decision and policy used, including why a receipt was not warranted.

## Decision 2: build the action manifest

Create a canonical, ASCII-safe or UTF-8 normalized manifest immediately before the side effect. It must contain:

- `agent_id`: stable identity of the acting Agent;
Read the full skill on GitHub

Details

Skill name
decision-receipts
Path
skills/decision-receipts/SKILL.md
License
MIT
Repo updated
8 September 2026

FAQ

What does the Decision Receipts skill do?

Use before or during consequential Agent actions such as deploy, delete, pay, export, access grants, or high-risk decisions when later offline verification of what happened and under which policy matters.

How do I install the Decision Receipts skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add sandbaseai/workbuddy-skill --skill decision-receipts. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Decision Receipts popular?

It lives in sandbaseai/workbuddy-skill, which has 9 GitHub stars and 3 forks.

Related skills

Browse the most popular Claude skills, all devops & cloud skills, or write your own with the skills generator.