plan-cache-rollout-gate

Aggregate plan-cache metrics across pipelines and emit a JSON rollout verdict.

1|Updated Mar 15, 2026
One-click install
npx skills add https://github.com/paulingham/.claude --skill plan-cache-rollout-gate
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: plan-cache-rollout-gate
Source: https://github.com/paulingham/.claude/tree/main/skills/plan-cache-rollout-gate
Command: npx skills add https://github.com/paulingham/.claude --skill plan-cache-rollout-gate

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Aggregates plan-cache metrics to determine if the rollout gate can flip from shadow to on, removing guesswork from critical deployments.

Core Features & Use Cases

  • Reads plan-cache metrics from hooks/plan-cache-audit.sh outputs and aggregates them across the larger window of last 30 pipelines or last 14 days, whichever is larger.
  • Applies three thresholds to decide verdicts: ROLLOUT_GATE_PASS, ROLLOUT_GATE_FAIL, INSUFFICIENT_DATA, and provides a structured payload for PR descriptions.
  • Designed to gate the flip-to-on PR with evidence, ensuring safe, observable rollout decisions.

Quick Start

Run the rollout gate against your metrics root to obtain a VERDICT payload for the flip-to-on PR.

Frequently Asked Questions about plan-cache-rollout-gate

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

FAQPage Schema
How do I evaluate plan-cache metrics before flipping a rollout gate from shadow to on?

To evaluate plan-cache metrics before a rollout, you aggregate data across the larger window of the last 30 pipeline sessions or 14 days. This applies three thresholds to emit a JSON verdict with evidence for your flip-to-on PR.

What is a data-driven rollout gate for pipeline deployments?

A data-driven rollout gate aggregates plan-cache metrics across pipeline sessions to decide if a feature can flip from shadow to on. It removes guesswork by applying threshold checks and providing a structured JSON verdict for evidence-based deployments.

How do I aggregate pipeline session data for a safe feature rollout?

You aggregate pipeline session data by reading plan-cache audit outputs and evaluating them over the larger of the last 30 sessions or 14 days. This produces a ROLLOUT_GATE_PASS, FAIL, or INSUFFICIENT_DATA verdict for your PR.

When do I need to check rollout gate thresholds for plan-cache deployments?

You need to check rollout gate thresholds before authoring the flip-to-on PR. Checking thresholds over the larger of 30 pipeline sessions or 14 days ensures you have sufficient data to prevent unsafe, unobservable rollout decisions.

What does an insufficient data verdict mean for plan-cache rollout decisions?

An INSUFFICIENT_DATA verdict means your plan-cache metrics did not meet the minimum window requirement of 30 pipeline sessions or 14 days. You must wait for more pipeline data before attempting to flip the rollout gate from shadow to on.

Can I use plan-cache rollout gate metrics without 30 pipeline sessions?

You can use the rollout gate metrics, but applying the gate requires the larger window of the last 30 pipeline sessions or 14 days. If you lack this volume, the aggregation will emit an INSUFFICIENT_DATA JSON verdict instead of a pass or fail.