Nm Imbue Proof Of Work

Enforce validation and evidence before declaring work complete. Use for acceptance criteria and done gates

This skill is an instruction-only checklist/framework for producing and capturing verification evidence — it does not install code or ask for secrets. However, when used an agent will be instructed to run local commands and read files and environment variables; the captured evidence may include sensitive data (process lists, config files, env vars, logs, tokens present in files, etc.). Before installing or running: (1) run the skill in a safe/sandboxed environment if you are concerned about leaking secrets; (2) review and sanitize any files or environment variables you don't want captured; (3) be aware the modules say proof:* evidence should be retained as audit trail (so logs may persist); (4) if you do not want the agent to run autonomously, restrict invocation or require user confirmation when the agent attempts to execute verification steps. If you want, I can list exact commands the SKILL.md suggests the agent will run so you can assess what outputs might be captured.

Type: OpenClaw Skill Name: nm-imbue-proof-of-work Version: 1.0.0 The skill bundle implements a 'Proof of Work' methodology designed to enforce technical verification and evidence-based reporting for AI agents. It includes comprehensive modules for TDD/BDD enforcement (iron-law-enforcement.md), structured evidence logging (evidence-logging.md), and anti-cargo-cult protocols (anti-cargo-cult.md). The instructions and examples use standard development tools (e.g., pytest, npm, jq, ps) to validate configurations and implementations. No indicators of data exfiltration, malicious execution, or harmful prompt injection were found; the bundle's logic is entirely focused on improving engineering rigor.