outcome-review

Measures shipped features against intake metrics and renders persevere, iterate, or kill verdicts.

Updated May 21, 2026
One-click install
npx skills add https://github.com/CagesThrottleUs/private-ai-harness --skill outcome-review-cagesthrottleus
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: outcome-review
Source: https://github.com/CagesThrottleUs/private-ai-harness/tree/main/skills/outcome-review
Command: npx skills add https://github.com/CagesThrottleUs/private-ai-harness --skill outcome-review-cagesthrottleus

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Teams write success metrics at intake but rarely verify whether shipped features actually moved those numbers, leaving launch success as an unverified claim. This Skill closes the measurement loop by comparing realized post-launch data against the north-star and input metrics promised at business-context-intake. ## Core Features & Use Cases - Metric Verification: Reads the business-context document's north-star and input metrics, pulls realized values from cited re-runnable sources (dashboards, queries, analytics events), and renders moved / did-not-move / inconclusive verdicts per metric. - HEART-Aware Timing: Applies Google HEART framework timelines so metrics are not declared failed before their earliest meaningful read (1-2 sprints for engagement, 1-3 cohorts for retention). - Launch Decision: Rolls metric verdicts up into a Lean Startup persevere / iterate / kill decision and feeds it back to portfolio WSJF re-ranking. - Use Case: A checkout feature shipped 32 days ago with a promised 85% completion target. Run this Skill to pull the actual completion rate from the analytics dashboard, check PR-FAQ assumptions, and record an evidence-backed persevere or iterate decision. ## Quick Start Run the outcome-review skill on the feature that shipped last month to check whether its north-star metric moved against the intake target.

Frequently Asked Questions about outcome-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I measure whether a shipped feature met its success metrics?

Run outcome-review after launch: it reads the business-context document's north-star and input metrics, pulls realized values from the cited dashboards or analytics events, and renders a moved, did-not-move, or inconclusive verdict per metric.

What is a north-star metric review after launch?

It is a post-launch check comparing the realized north-star metric against the baseline and target set at intake. This Skill automates that comparison and rolls the results into a persevere, iterate, or kill launch decision.

When is it too early to declare a metric failed?

Per the HEART framework timelines, task success and engagement metrics need 1-2 sprints, while retention and happiness need 1-3 cohorts (about 1-3 months). Before those windows, the Skill records inconclusive-too-early instead of a failure verdict.

What happens if a metric has no data source?

A metric without a re-runnable source such as a dashboard URL, query, or analytics event is recorded as unmeasurable-as-specified, never as met. The gap is routed back to observability-standards for instrumentation.

When should I skip outcome review?

Skip it for bug fixes, refactors, config changes, and internal-only work that have no business-context north-star metric. If no .ai/<feature-slug>/business-context/ document exists, there is nothing to measure against.