Abhillashjadhav

2026 08 27 Complete Ai Evals Suite

Name: 2026 08 27 Complete Ai Evals Suite
Author: Abhillashjadhav

AI community

SOURCE https://github.com/Abhillashjadhav/AI-PM-essential-skills/blob/main/prompts/2026-08-27-complete-ai-evals-suite.md

Description

# Complete AI Evals for PMs as a full evaluation suite ## Objective Extend the existing `pm-verifier` package and CLI into a complete, fail-closed evaluation suite. Preserve the public product name **AI Evals for PMs**, the stable technical identifier `pm-verifier`, all existing commands, and all existing outcome/trajectory behavior. Do not create a second product, repository, plugin, or CLI. ## Problem and hypothesis The current harness can grade supplied outcomes and trajectories, but a P

Installation

Installs to ~/.claude/skills/abhillashjadhav-ai-pm-essential-skills-2026-08-27-complete-ai-evals-suite/SKILL.md

Terminal

mkdir -p ~/.claude/skills/abhillashjadhav-ai-pm-essential-skills-2026-08-27-complete-ai-evals-suite && curl -fsSL https://raw.githubusercontent.com/Abhillashjadhav/AI-PM-essential-skills/HEAD/prompts/2026-08-27-complete-ai-evals-suite.md -o ~/.claude/skills/abhillashjadhav-ai-pm-essential-skills-2026-08-27-complete-ai-evals-suite/SKILL.md

Restart Claude Code, or start a new session, for it to be picked up.